| 1 | ChatGPT | OpenAI | America | Interactive AI, text generation, summarization, translation, programming support | Multilingual support | Free/Plus ($20/month) | 3.1 billion visits per month | Widely adopted by educational institutions and companies | https://chat.openai.com | DetailsChatGPT is a conversational AI chatbot released by OpenAI in November 2022, and is an advanced natural language processing model based on the GPT (Generative Pre-trained Transformer) series. The initial version, GPT-3.5, enabled natural and contextual interaction with users and quickly gained attention from developers and general users around the world. In March 2023, an even higher-performance GPT-4 will be released, dramatically improving accuracy and logical reasoning ability. Research institutions and companies are using it in a wide variety of ways, including summarizing papers, automating customer correspondence, and supporting the creation of teaching materials in educational settings.
Technically, it uses a transformer architecture and performs self-supervised learning using large-scale data, followed by instruction-based fine tuning to strengthen its ability to "follow instructions." A big appeal is that you can link with external tools through plug-ins (web browsing, code execution, database search, etc.) and add functions such as acquiring the latest information, file manipulation, and spreadsheet calculations. Additionally, since it can be incorporated into proprietary applications via API, it is used not only for chatbots but also as the core of customer support and business efficiency tools.
There are countless use cases, ranging from business document drafting to marketing copy generation, program code debugging support, and even real-time translation between multiple languages such as Japanese and English. It contributes to productivity improvements in a variety of areas, including dialogue engines for knowledge sharing platforms within companies, automatic grading tools for educational institutions, automatic responses to FAQs, and assistance with paper reviews by researchers.
On the other hand, concerns have also been raised about the issue of "hallucination," in which the output content is not necessarily accurate, and privacy/security concerns. Although model updates are gradually reducing these risks, final checks by experts are essential for important decisions. In addition, it is necessary to design a plan that takes API costs and response speeds into consideration, and total cost management is the key, especially when implementing on a corporate scale. |
| 2 | Canva | Canva Pty Ltd | Australia | Graphic design, presentation creation | Multilingual support | Free/Pro ($12.99/month) | 841.3 million monthly visits | Popular in marketing and education fields | https://www.canva.com | DetailsCanva is an online platform that allows anyone to create professional-quality graphics without any design knowledge. Founded in Australia in 2013, you can intuitively design posters, social media posts, presentation slides, business cards, flyers, and more with just drag-and-drop operations. There are over tens of thousands of templates available, and a wide variety of fonts, icons, and photo materials are available even with the free plan. The Pro plan unlocks more advanced features such as brand kit functionality, team collaboration, video editing, and export of transparent background images.
On the technical side, it is equipped with an internal engine that automates image processing and layout calculations, and AI supports automatic layout adjustment and color scheme suggestions based on the user's intentions. From 2024 onwards, AI text generation functions such as "Canva Magic Write" and automatic image editing AI (background deletion, style conversion, etc.) will be integrated, further expanding the range of content creation. It also has powerful team collaboration through joint editing and commenting functions, and has a wide range of users, including marketing departments, educational institutions, and individual creators.
Canva's main use cases include creating banners for social media, visually enhancing sales materials, creating slides for online courses, and generating eye-catching images for blog posts. For companies, use Brand Kit to centrally manage corporate logos and colors and share internal templates to streamline brand governance. We also offer Canvas for Education, a plan for educational institutions that supports the creation of learning materials, and is actively used in remote classes and workshops.
Our strengths are a UI design that anyone can use, a wealth of materials and templates, and a seamless cloud-based sharing function. However, there are some areas that are lacking for professional designers who seek detailed customization, and in particular, advanced vector editing and fine-tuning of details are typically used in conjunction with specialized tools. That said, the ability to mass-produce high-quality designs in a short time at low cost is a great advantage whether it's for business, education, or hobbies. |
| 3 | DeepL | DeepL GmbH | Germany | Highly accurate machine translation | Supports 31 languages | Free / Pro (from ?7.49/month) | 303.5 million visits per month | Evaluated for high translation accuracy | https://www.deepl.com | DetailsDeepL is a machine translation service developed by DeepL GmbH, which was founded in Cologne, Germany in 2017, and its greatest feature is high accuracy that surpasses conventional NMT (neural machine translation). Especially in major languages such as Japanese, English, German, and French, it is possible to provide natural translations based on the context, and has received high praise for its readability and consistency compared to industry standards such as Google Translate and Microsoft Translator. In addition to easily translating using a web browser, it also supports desktop apps for Windows/macOS, MS Office integration, and system integration through API, making it flexibly adaptable to business translation workflows.
DeepL's translation engine combines a unique Transformer-based architecture with massively parallel processing, and is fine-tuned using example datasets and specialized field corpora. It has a style guide function that allows users to specify tone such as "technical terminology," "honorific language," and "casual expressions," and can also handle domain-specific projects such as translation of marketing documents, technical manuals, and contracts. With the Pro plan, you can customize terminology, save translation memory, and upload entire files for batch translation.
Use cases are wide-ranging, including multilingualization of websites, translation of product descriptions on e-commerce sites, global expansion of in-house materials, translation support for research papers and patent documents, and preparation of presentation materials for international conferences. An increasing number of companies are building hybrid workflows that combine automatic translation and manual proofreading, and this has become established as a means to reduce translation costs and shorten delivery times.
On the other hand, there are cases where professional services are preferred for more niche language pairs and real-time conversation interpretation functions, and DeepL is expected to strengthen its support in the future. In addition, in terms of data security, there are calls for higher guarantees such as the use of private clouds in the Pro plan, and the challenge is to enhance enterprise support. |
| 4 | Google Gemini（旧称Bard） | Google | America | Search support, text generation, multimodal support | Over 40 languages | Free / Gemini Advanced ($20/month) | 267.8 million monthly visits | Strength is cooperation with Google services | https://gemini.google.com | DetailsGoogle Gemini (formerly known as Bard) is a conversational AI platform developed by Google's cutting-edge AI research team. Announced as a beta version in February 2023, it is good at accessing the latest information and checking facts by integrating with large-scale search indexes and knowledge graphs that the company has accumulated over many years. Multiple models such as Gemini Ultra/Pro are available, and it has excellent logical reasoning ability and multimodal (text, image, audio) conversion performance, and can not only generate text but also create image captions and data visualization summaries.
Technically, in addition to the Transformer, it is equipped with a multimodal dedicated encoder/decoder network. Its major strength is that it can be linked with web search results, Google Docs, and spreadsheets, allowing you to respond while referring to the latest research papers and news articles in real time. Integration through API is also progressing, and more and more companies are incorporating Gemini into customer support chats, product recommendations, data analysis dashboards, etc.
Use cases include improving the efficiency of research work (market research, competitive research), generating drafts for reports and proposals, automatically generating image descriptions, and automatically responding to in-house help desks. In particular, it has been praised for its high compatibility with Google search and the ability to obtain evidence-based answers (with links to sources), making it safe to use for business decision-making and academic research research.
Challenges include reducing speed when access is concentrated, data handling policies when handling highly confidential information, and strengthening measures against the "hallucination" risk that is common to all generative AI. In the future, we expect to expand the provision of private models on Google Cloud and the expansion of dedicated environments where in-house data can be safely studied. |
| 5 | Character AI | Character.AI | America | Customizable interactive AI character creation | English Center | free | 220.2 million visits per month | Popular in the entertainment field | https://beta.character.ai | DetailsCharacter AI is a platform where users can define their own characters (personalities) and enjoy conversations based on those settings. It will be released in beta at the end of 2022, and you can easily create and share a wide variety of "AI partners" such as movie and game characters, historical figures, and original characters. Each character behaves dynamically based on prompts given by the user, such as personality, tone, range of knowledge, hobbies and preferences, and realizes natural dialogue that includes humor and emotional expression.
The backend uses a large-scale language model (combined with Anthropic and OpenAI models) to sequentially generate output based on user-defined prompt templates. By observing conversations between characters and participating in public character communities, you can also experience advanced dialogue scenarios created by other users. Commenting and rating features on the platform facilitate more natural and immersive character creation.
For educational purposes, it can be used as an "interactive learning tool" for discussions with historical figures, or as a companion for English conversation learning. In the entertainment field, it is also popular as a source of inspiration for creating interactive stories with original characters and writing scripts. Additionally, the number of cases where companies are turning brand characters into AI and using them as user-friendly chatbots is increasing.
On the other hand, since it depends on character settings, it is necessary to improve the accuracy of the safety filter that controls unintended speech and behavior. In particular, it is important for administrators to monitor and educate users in order to prevent the spread of statements that violate public order and morals and fake information. In addition, since it is an English-based platform, Japanese support is still in the beta stage. It is expected that multilingual support will be strengthened in the future. |
| 6 | Grammarly | Grammarly Inc. | America | English grammar check, writing improvement support | English only | Free/Premium ($12/month) | 37.7 million visits per month | Standard for students and business people | https://www.grammarly.com | DetailsGrammarly is a web service and application that helps improve the quality of English writing, and was founded in the United States in 2012. It detects grammatical and spelling errors in real time, suggests appropriate corrections, and provides extensive feedback on writing style (formal/informal), vocabulary diversity, redundancy reduction, and readability score. It has a variety of clients such as Chrome extensions, Word add-ins, desktop apps, and mobile apps, and is widely used for email creation, reports, SNS posting, and paper writing.
On the technical side, we use a unique engine that combines natural language processing and machine learning, and pre-train with a corpus of hundreds of billions of words. By performing sentence structure analysis and style analysis, we propose not only grammar but also word substitutions and expression improvements based on the context. With the Premium plan, you can also use genre-specific style guide settings and plagiarism checker, which enhances quality control when creating academic papers and business documents.
Use cases are wide-ranging, including ensuring the quality of business communication documents, supporting the writing of English articles by non-native writers, language proofreading of academic papers, brushing up resumes for job hunting, and creating catchphrases for SNS content. Especially for English learners, real-time explanatory feedback is popular as a tool to improve learning effectiveness.
On the other hand, it is difficult to deal with specialized field-specific terminology and writing styles, and there is a risk of incorrect corrections when translating and proofreading technical and legal documents. In addition, since cloud processing is not completed in a local environment, it is necessary to establish an implementation policy when handling highly confidential documents. Recently, the development of AI translation functions and document summarization functions has progressed, and the platform continues to become more multifunctional. |
| 7 | Microsoft Copilot | Microsoft | America | AI support within Office products (writing, data analysis) | Multilingual support | Included with Microsoft 365 subscription | Implemented by over 37,000 organizations | Highly rated as a productivity improvement tool | https://www.microsoft.com/en-us/microsoft-365/copilot | DetailsMicrosoft Copilot is an AI assistant integrated into Microsoft 365 applications such as Word, Excel, PowerPoint, and Outlook. Launched in March 2023, it will dramatically streamline business workflows, including automatically generating and summarizing Office documents, extracting insights from spreadsheet data, and drafting presentation materials. For example, you can generate the outline of a proposal or report just by giving a prompt in Word, and in Excel, if you tell it in natural language to ``analyze monthly trends from sales data,'' pivot tables and graphs will be automatically created.
Copilot's AI engine is based on OpenAI's GPT-based model, and works with Microsoft Graph (user organizational data) and Azure AI services. It features personalized suggestions based on context such as the user's email, calendar, and document history. Rather than directly reusing in-house data, the architecture emphasizes data security and privacy protection, and also provides administrator settings and audit log functions.
Case studies include a major manufacturing company using it to automatically create quality reports, and a consulting firm using it to create drafts of market research reports. There are reports that even small businesses have reduced the number of man-hours required to create sales proposals by an average of 30%. Further expansion to Dynamics 365 and Power Platform is planned in the future, and AI assistance in low-code development environments is also expected to accelerate. |
| 8 | Perplexity AI | Perplexity AI Inc. | America | AI search engine features answers with information sources | Mainly English (Japanese beta) | Free / Pro ($20/month) | 10 million monthly active users | Specializing in providing accurate information | https://www.perplexity.ai | DetailsPerplexity AI is a next-generation search engine that uses AI to generate summary answers to search queries and reveals the underlying information sources (web pages, papers, news articles, etc.). Launched in 2023, it differs from traditional keyword matching searches by understanding questions in natural language, cross-referencing multiple sources, and providing highly reliable answers. Since a list of links and cited documents are shown along with the answer, it is easy to see at a glance what information led to that conclusion.
Technically, it uses a large-scale language model and collaborates with real-time web crawling information and third-party databases. With the Pro plan, you can improve response speed with priority access and integrate internal systems with API. There is also a free-form discussion mode and a history management function, allowing you to build a workflow that completely leaves part of your research work to AI.
Use cases include creating market research reports, doing preliminary research for legal and regulatory checks, reviewing medical literature, and obtaining summaries of technology trends. In the education field, it is used as a tool for preparing study materials for teachers and for supporting report writing for students. Answers with evidence are especially useful in situations where logical support is required, and are increasingly being introduced in corporate compliance departments and research and development departments. |
| 9 | Claude | Anthropic | America | Natural conversation, good at processing long texts, emphasis on safety | English Center | Free / Pro ($20/month) | Available in over 159 countries | Emphasis on safety and ethics | https://www.anthropic.com | DetailsClaude is a large-scale, safety-focused language model developed by Anthropic. It was released in beta as "Claude 1" in 2023, and after passing through the next generation "Claude 2" and "Claude 3", the latest model appeared in 2024 with significantly improved dialogue quality, long text summarization, and reasoning ability. It is characterized by the introduction of Anthropic's unique "stakeholder alignment" method and emphasis on safety guardrails that suppress the generation of false information and harmful output.
Claude is strong against long text input and can consistently summarize and edit documents with tens of thousands of tokens. In multi-turn dialogue, it can accurately trace the flow of conversation and has excellent recall ability (the ability to remember previous statements). For businesses, custom tuning via API and support for implementation in on-premises environments are available, allowing operations to comply with financial regulations and medical regulations.
Specific usage examples include legal document summaries and interactive reviews, drafting of research reports, customer support chatbots, and internal knowledge base autoresponders. Particularly in the medical field, it is currently being introduced to summarize academic papers and support clinical data analysis. PoC (proof of concept) is also being conducted at government agencies and major financial institutions that have strong security concerns. |
| 10 | Jasper | Jasper AI | America | Marketing text generation and content creation support | English Center | Free trial available / Paid plan | Number of monthly users not disclosed | Popular with marketers | https://www.jasper.ai | DetailsJasper is a text generation AI platform specifically focused on creating marketing content. Released in 2019 as “Jarvis” and renamed to “Jasper” in 2021. Equipped with many templates optimized for each purpose, such as blog articles, social media posts, ad copy, email campaigns, and the Hero section of landing pages, even non-writers can mass-produce high-quality text in a short time. If you use the built-in "Boss Mode", you can support everything from structural design of long articles to draft generation.
Jasper's engine is based on OpenAI's GPT model, but it has been fine-tuned independently and learned marketing terminology and sales writing techniques. It has a wide range of functions for marketers, such as generating SEO-optimized articles from keyword input, suggesting CTA (Call To Action) wording, and generating wording variations for A/B testing. Plugins for linking with HubSpot, Surfer SEO, and Google Docs are also provided, making it easy to integrate into existing workflows.
Usage scenarios include e-commerce product descriptions, newsletters, press release drafts, YouTube video descriptions, and landing page copy production. Multilingual support is mainly in English, but support for Japanese has also started in the beta stage, and future expansion is expected. Team collaboration features and approval workflows are also available, making it suitable for group use at agencies. |
| 11 | DALL・E3 | OpenAI | America | Image generation from text | English (prompted input) | Free (limited) / Credit system | Number of monthly users not disclosed | Popular with creators | https://openai.com/dall-e | DetailsDALL·E is an AI model developed by OpenAI that generates high-quality images from text, and is a product with particularly excellent creativity and visual expressiveness among generative AI technologies. The first version was released in 2021, and in 2023, DALL·E 3 arrived, allowing users to better understand their text prompts and generate precise, artistic visuals. In addition to simply expressing words such as "cat" or "outer space," complex scenes such as "a cat wearing a space suit eating pizza on the moon" can be faithfully expressed.
Technically, we use a transformer architecture that combines a text encoder and an image decoder, and perform pre-training on a large-scale text/image pair dataset. DALL·E 3 reproduces image composition, color, texture, perspective, etc. with high precision, and flexibly supports post-generation editing and variation creation using Inpainting and Variations functions.
The main use cases are wide-ranging, including eye catches for SNS, visual materials for blogs, advertising creatives, concept art for online games, and illustrations for educational materials. Integration with OpenAI's ChatGPT allows image generation while interacting on the chat interface, making it highly adaptable to creative workflows.
Challenges include copyright issues with generated images and risk management of fake images, and OpenAI is working to improve safety by strengthening filtering of inappropriate content. When using for commercial purposes, it is necessary to check the terms of use and credit system. |
| 12 | Midjourney | Midjourney Inc. | America | High-quality image generation AI | English (prompted input) | $10?60/month | 2.8 million visits per month | Support for artists and designers | https://www.midjourney.com | DetailsMidjourney is a Discord-based high-quality image generation AI platform provided by Midjourney Inc., and has been gaining support from the creative industry and individual creators since its beta version was released in 2022. Users can generate four candidate images in seconds to tens of seconds by simply entering the /imagine command on a Discord server and sending a text prompt. After generation, you can use the "upscale" and "variation generation" functions to adjust the details and create the visual you want.
On the technical side, the Transformer-based model is trained in advance using large-scale text and image data, and light source expression and texture reproduction have been significantly improved since version 5. Since you can specify seed values and CFG scales, you can adapt to various art styles by devising prompts and adjusting parameters. Feedback and sharing functions are active in the community channel, and its strength is that it allows users to exchange technical and expressive know-how.
Example use cases include concept art for game and video production, marketing materials, SNS posts, website visuals, and print design. Plans include Basic ($10 worth of credits per month), Standard, and Pro, and credit consumption and compliance with guidelines are required for commercial use. In the future, we plan to provide on-premise models and a custom model training function for enterprises. |
| 13 | Runway ML | Runway | America | Video editing, AI video generation | English Center | Free (limited) / $12/month? | 1.3 million visits per month | Support for video creators | https://runwayml.com | DetailsRunway ML is a cloud-based AI video editing and generation platform provided by Runway, which has brought innovation to the video production scene since its beta release in 2020. Key features include Gen-2, which generates videos from text, Reimagine, which applies new styles and effects to existing videos, automatic transition generation, background replacement, and motion tracking, allowing you to create professional-quality videos using just a GUI on your browser.
All video processing is performed on high-speed GPUs on the cloud, so it can be used without depending on the specifications of the local machine. Plug-ins for Figma, Adobe After Effects, and Premiere Pro are also provided, making it easy to integrate with existing workflows. It is suitable for rapid prototyping of marketing videos, educational content, and social media short clips, and is increasingly being adopted by small and medium-sized creative teams.
There are three tiers of plans: Free (with limited functionality), Starter ($12 worth of credits per month), and Pro. Pro unlocks priority processing, high-resolution export, and API access. Please note that AI-generated video may occasionally contain noise or repetitive patterns, and manual adjustments may be required for final output. In the future, we plan to enhance enterprise functions such as real-time collaborative editing and model provision in an on-premises environment. |
| 14 | Poe | Quora | America | A chat platform that can use multiple AIs (ChatGPT, Claude, etc.) at once | Multilingual support | Free / Premium ( | | | | DetailsPoe is a service provided by Quora and is an AI platform that allows you to use multiple large-scale language models such as ChatGPT, Gemini, and Claude from a single interface. Users can switch between models with one click and compare answer quality and speed for the same question. The free plan also provides ample answer credits, making it ideal for the early phases of research and brainstorming. Dialogue history is saved in thread format, making it easy to refer to and reuse later. Plug-ins allow you to quickly adapt to the latest LLM version, and you have the advantage of always being able to try out the latest models. You need to be careful about credit consumption, and it is recommended to consider a paid plan when using a large amount. |
| 15 | DeepSeek | DeepSeek-AI | China | Chat AI with advanced reasoning ability | English, Chinese | Free / Premium plan available | 268 million monthly visits | Attract attention with advanced reasoning ability | https://deepseek.com | DetailsDeepSeek is an enterprise platform that integrates document analysis and natural language search. Its biggest feature is that it centrally indexes various sources such as PDF, Word, and web pages, and allows you to instantly extract the necessary information "just by asking questions in natural language." Specifically, it provides a wealth of functions that directly lead to increased operational efficiency, such as full-text search and summarization of patent documents and research papers, automatic FAQ generation, and risk highlights for compliance documents. The Pro plan allows you to add a dedicated API and custom terminology dictionary, and also supports company-specific technical terms. In terms of security, it supports OAuth and SSO collaboration, and is also expected to be used in confidential data environments. |
| 16 | JanitorAI | JanitorAI | America | Customizable interactive AI character creation | English Center | free | 200 million visits per month | Popular in the entertainment field | https://janitorai.com | DetailsJanitorAI is a platform that allows you to build custom chatbots with no code. When you drag and drop Google Sheets, Notion, or PDF files to upload, the bot automatically creates a knowledge base and uses it for conversation. You can deploy FAQ response and customer support bots by simply configuring the ``Introduction,'' ``Knowledge Base,'' and ``Guardrails'' on the GUI, and can easily embed them on corporate sites and internal portals. You can try the basic functions with the free plan, and the business plan allows you to increase the number of simultaneous sessions and obtain detailed logs. |
| 17 | Remove.bg | Kaleido AI | Austria | Automatic image background removal tool | Multilingual support | Free (with limitations) / Paid plans available | 112 million monthly visits | Popular with designers and photo editors | https://www.remove.bg | DetailsRemove.bg is an automatic image background removal service powered by Kaleido AI that makes the background of people and objects transparent with one click. We implement a deep learning-based segmentation model to extract details from hair and animal fur with high precision. We provide plug-ins and APIs for Photoshop and Figma, and support large-scale workflows such as batch processing of e-commerce product images and mask generation for AR apps. The free plan is limited to 50 images per month, and the Pro plan and higher offers high-resolution output, batch processing, and support with an SLA. |
| 18 | Duolingo | Duolingo Inc. | America | Language learning app using AI | Multilingual support | Free / Super Duolingo ($6.99/month) | 96 million visits per month | Improving learning effectiveness with gamification | https://www.duolingo.com | DetailsDuolingo is a language learning platform that combines gamification and AI chatbots. The lessons are basically stage-clearing type, but the AI Duo mode implemented in 2023 allows daily conversation simulation, and the Roleplay function allows you to practice dialogue by role, such as in travel or business situations. Learning progress is visualized on the dashboard, and AI analyzes weak patterns and automatically generates a review plan. The Super Duolingo plan offers no ads, additional exercises, and unlimited AI chat, making it widely adopted by both educational settings and individual learners. |
| 19 | CapCut | Bytedance | China | AI-powered video editing app | Multilingual support | free | 85 million visits per month | Popular with TikTok users | https://www.capcut.com | DetailsCapCut is an AI-powered video editing app provided by Bytedance. It has mobile and PC versions, and allows you to apply effects, select templates, and auto-cut with tap operations or drag and drop. AI extracts highlights through automatic scene analysis, and also features automatic subtitle generation and beat synchronization with music. It significantly reduces the man-hours required to create short video content, and the template sharing community is also active. Although you need to be careful about PC specifications when editing 4K, the barrier to entry is low for free, and its use is rapidly increasing, especially among TikTok creators. |
| 20 | ElevenLabs | Eleven Labs Inc. | America | High-quality speech synthesis and cloning | English Center | Free trial available / Paid plan available | Number of monthly users not disclosed | Used for narration and podcast production | https://elevenlabs.io | DetailsElevenLabs is a natural speech synthesis platform that applies WaveNet-based models, allowing you to fine-tune voice quality, intonation, speed, and emotional tone. Voice Lab allows you to clone a speaker from a few minutes of audio sample and generate a new text voice with the same voice actor. Use cases are wide-ranging, including audiobook production, podcasts, game narration, and anime character voices, and the API and SDK allow for easy application embedding. After the free trial, you will be charged for the usage time, and a license agreement and voice actor consent are required for commercial use. |
| 21 | Gamma | Gamma App | America | Presentation creation tool using AI | English Center | Free / paid plans available | Number of monthly users not disclosed | Contributes to more efficient slide creation | https://gamma.app | DetailsGamma is a cloud-based presentation creation tool that utilizes AI. Just by selecting keywords and templates, you can automatically create everything from storyline generation to slide division, headings/bulleted text, and simple graph generation. When you register brand guidelines and logos, they are automatically reflected, and the team sharing function also supports collaborative editing. You can try the basic functions with the free plan, and the Pro plan and above allows API integration and custom theme creation. |
| 22 | Notion AI | Notion Labs Inc. | America | Added AI support to document creation and management | English Center | Included with Notion subscription | Number of monthly users not disclosed | Evaluated as a productivity improvement tool | https://www.notion.so/product/ai | DetailsNotion AI is an AI assistant function that allows you to receive various support such as summarizing, creating drafts, translating, generating ideas, and extracting key points from table data by entering the “/ai” command at any point in a Notion document. It is possible to maintain a personal style by learning the user's past documents and output a composition proposal according to a template. With the corporate plan, you can turn on and off the AI function for each workspace, and you can also flexibly customize settings for translation quality and generation accuracy. |
| 23 | HubSpot Email Writer | HubSpot | America | Email creation support tool using AI | English Center | Included with HubSpot subscription | Number of monthly users not disclosed | Popular with marketers | https://www.hubspot.com/products/marketing/email | DetailsHubSpot Email Writer is an AI-powered email writing experience integrated into HubSpot's marketing and sales platform. Users can automatically generate and edit subject lines and body text using the forward slash (/) and highlighting features, allowing users to create hundreds of sales email drafts in minutes. The created text is seamlessly linked with HubSpot's CRM, and analytical data such as open rate, click rate, and response rate can be automatically obtained and visualized, speeding up the cycle of verifying the effectiveness of measures and making improvements.
The AI engine learns past sending history and response data accumulated in the customer database, and automatically makes personalized recommendations tailored to customer segments (industry, job title, region, etc.). In addition, it centrally supports a variety of use cases in email marketing, such as ``optimal tone and structure for each segment,'' ``generation of variations for A/B testing,'' and ``automatic generation of subject line candidates.''
Emails you create can be operated not only on your desktop but also on the mobile app, allowing you to draft, send, and track emails on the go. Another great feature is that by integrating with existing email clients such as Gmail and Outlook, you can send and receive emails with a native experience without leaving your HubSpot account.
In addition, HubSpot AI will work with functions as a marketing automation platform, such as ``automatic input of contact information from signatures,'' ``template management,'' and ``simultaneous distribution campaign settings,'' to improve the efficiency of overall marketing activities. The AI draft creation function is available even with the free plan, and if you wish to use more advanced analysis and custom models, you can upgrade to the paid Marketing Hub Professional/Enterprise plan for more detailed customization and SLA support.
Companies that have introduced it have reported results such as ``reducing initial creation time for large drafts by more than 90%,'' ``increasing open rate by 20%,'' and ``improving campaign ROI by an average of 1.5 times,'' and it has been particularly well received by B2B sales organizations and SaaS companies in their outreach operations. On the other hand, since the function design is centered around English, there are some parts that are not compatible with Japanese localization, and post-editing may be required for important customer journey materials. In the future, it is expected that AI models and templates for Japanese will be enriched. |
| 24 | Reclaim | Reclaim AI | America | AI-based schedule optimization tool | English Center | Free / paid plans available | Number of monthly users not disclosed | Contributes to more efficient time management | https://reclaim.ai | DetailsReclaim is a smart calendar that seamlessly syncs with Google Calendar and Outlook and uses AI to automatically optimally arrange tasks, habits, meetings, and breaks. Users can set "Habits (regular events)" and "Tasks (single smart events)" and specify deadlines and priorities, and AI will detect free time and automatically insert them into the optimal slot. Team Plan optimizes each member's calendar cross-sectionally and automates the coordination of 1on1 and team meetings. By using the scheduling link function, you can change the schedule that is offered as "free time" up or down according to priority, and book highly important meetings with priority.
In addition, the Buffer Time function automatically inserts rest time (decompression time) after meetings or before important tasks, and travel time. Weekly performance reports help optimize productivity and work-life balance by visualizing ``secured focus time,'' ``protected break time,'' and ``analysis of travel time.'' It has been installed by over 60,000 companies and over 500,000 users, and received a 4.8/5 rating in G2. Companies that have introduced it have reported results such as a 30% or more reduction in meeting coordination costs, a 20% increase in focus time, and a reduction in burnout syndrome. On the other hand, some say that detailed task management functions (subtasks and note functions) are lacking, so heavy users are recommended to use a dedicated task management tool. |
| 25 | ClickUp | Mango Technologies | America | AI-powered project management tool | Multilingual support | Free / paid plans available | Number of monthly users not disclosed | Contribute to improving team productivity | https://clickup.com | DetailsClickUp Brain is a suite of role-based and context-aware AI assistants built into the ClickUp platform. Access any thread of a task, document, or conversation and use natural language prompts to summarize, extract next actions, automatically create progress reports, generate templates, insert snippets, and more. Over 100 predefined templates are available for each role (PM, Engineer, Marketer, etc.) to facilitate automation that fits your team's workflow. Furthermore, it facilitates asynchronous communication by analyzing conversation logs and documents, instantly extracting key points, and displaying the TL;DR of long comments.
Also available from the IDE or mobile app, the Write with AI feature generates drafts for emails, blogs, and copywriting in seconds. It also supports integration with Google Workspace, GitHub, and Slack, and covers a wide range of use cases such as project outline drafts, test plans, QA creation, code reviews, and design document creation. Users have said things like ``reducing man-hours by more than 5 hours a week,'' ``reducing project startup time by 70%,'' and ``improving the quality of email text generation.'' Depending on the plan, the AI function is a paid add-on, so budget planning is required, but the hurdles to implementation are relatively low, and it has been widely adopted by small teams to large organizations. |
| 26 | Asana | Asana Inc. | America | Task/project management tool with AI functionality | Multilingual support | Free / paid plans available | Number of monthly users not disclosed | Widely adopted from large companies to small and medium-sized enterprises | https://asana.com | DetailsAsana AI provides features such as "Smart Status," "Smart Chat," "Smart Fields," "Smart Summaries," "Smart Rule Creator," and "Smart Goals" to support each phase of project management with AI. Achieves consistent workflow optimization, including triage of new work requests and automatic assignment of personnel, automatic generation of status reports based on task progress, natural language report creation instructions, and response to inquiries using a chatbot interface. AI-based resource allocation suggestions, risk detection, and knowledge base search functions are also integrated to support rapid data-driven decision-making.
In June 2024, we will launch the "AI Teammates" feature, which will allow you to delegate some tasks to AI. Users set triggers and rules, and the AI takes care of creating new project outlines, regular reports, and executing repetitive tasks. After conducting PoC at major companies such as Palo Alto Networks, we expanded to an enterprise plan. The design adopts a human-centered design in which final checks are performed by humans, and takes compliance requirements into consideration. On the other hand, since the UX design is mainly English-based, future challenges include localization support and strengthening Japanese support. |
| 27 | Gradio | Hugging Face | America | Machine learning model demo creation tool | English Center | Free / paid plans available | Number of monthly users not disclosed | Supported by researchers and developers | https://gradio.app | DetailsGradio is an open source library that allows you to quickly build web-based GUI demos for machine learning models and arbitrary functions with just a few lines of Python code. Using gr.Interface and gr.Blocks, we combine a variety of components such as text input, sliders, image uploads, and graph output, and immediately publish demos. Manage thousands of requests with SSE-based built-in queuing and support for one-click deployment to Hugging Face Spaces. It is used as a place for researchers and data scientists to quickly share model results with non-technical people, and has a track record of being ranked number one in GitHub trends.
With support for community-made components and custom CSS/JS, there is a high degree of freedom in design, and companies have adopted it for a variety of use cases such as model visualization, speech recognition demonstrations, and visualization of reinforcement learning agents. Installation is completed with pip install gradio, and it can be used from Jupyter Notebook, Colab, or a local server. It has received high praise from users, including ``reducing demo development time by 90%'' and ``shortening evaluation cycles through collaboration.'' |
| 28 | Synthesia | Synthesia Ltd. | England | AI video generation platform | English Center | Paid plan only | Number of monthly users not disclosed | Used for creating corporate training videos | https://www.synthesia.io | DetailsSynthesia is an AI video generation platform that turns text scripts into videos in over 140 languages. Select a narrator from over 230 types of AI avatars (customizable with company logo and brand colors) and generate realistic lip-sync synchronized audio and video based on the input text. In-Editor allows prompt operations such as script editing, tone changes, summarization, translation, etc., and supports embedding into LMS and major CMS.
For companies, we provide API integration, SAML authentication, and a company-specific avatar creation function (Digital Twin) to streamline the mass generation of training videos, product demos, and marketing content. We have strengthened security with SOC 2 Type II/GDPR compliance and have also rolled out Avatar Builder, which allows you to create a digital avatar in just a few minutes with a photo of your face. Used by over 50,000 companies and received a 4.7/5 rating from G2. In terms of funding, we achieved a valuation of $2.1 billion in 2025 and continue to see strong growth. However, it has been pointed out that there are limits to CG quality and delicate facial expression control, and in some cases it is necessary to use it in conjunction with conventional shooting for high-end advertising purposes. |
| 29 | AdCreative.ai | AdCreative.ai | France | Advertising creative generation using AI | English Center | Free trial available / Paid plan available | Number of monthly users not disclosed | Advocated by marketers | https://www.adcreative.ai | DetailsAdCreative.ai is an AI tool that automatically generates "high conversion rate visuals + copy" for SNS, search, and display advertising. Generate a variety of assets in seconds, including advertising banners, video ads, text/headlines, product photoshoots, full video ads, and more. Using our unique creative data set of over 1B and real-time performance analysis, we propose layouts, color schemes, and wording that aim to improve CTR and CVR.
Use the template builder function to create templates that reflect brand guidelines and mass-produce advertising materials that are consistent across the team. By linking advertising accounts, it is also possible to analyze past creatives and visualize competitive advertising benchmarks. As a result, it has been reported that "ad production time was reduced by more than 90%," "CVR increased by an average of 14 times," and "team productivity increased by 30%." You can start using the free plan, and higher-end plans offer advanced features such as "competitor insights," "creative scoring," and "personalized AI models." Multilingual support is currently being strengthened, and further localization functionality is expected to be expanded in the future. |
| 30 | Tabnine | Codota | Israel | Code completion and development support using AI | Multilingual support | Free / Pro ($12/month) | Number of monthly users not disclosed | Contributes to improving developer productivity | https://www.tabnine.com | DetailsTabnine is an IDE-integrated AI coding assistant with line-by-line and entire-function completion, code generation from natural language comments, test case generation, refactoring suggestions, and code explanations. The IDE Chat feature provides natural language instructions to the chat interface to support a wide range of tasks such as patch generation, bug fixing, test design, and documentation creation. AI models are trained on OSS licensed code with an emphasis on privacy and license compliance.
Provides plug-ins for major IDEs such as VS Code, JetBrains products, Visual Studio, and Eclipse. Build custom models shared by your team and get suggestions optimized for your project's unique codebase. Use cases include "initial automation of code reviews," "support for understanding legacy code," "expanding test coverage," and "automatic correction in CI/CD pipelines." A model selection function will be released in February 2024, and it will also support Mistral and OpenAI model switching. It has over 1 million users and raised $25 million in Series B in 2023. Enterprise implementations have reported a 20% improvement in development speed and a 15% improvement in bug fixing productivity. |
| 31 | DataRobot | DataRobot, Inc. | America | Machine learning model automation platform | English Center | custom plan | Number of monthly users not disclosed | Accelerate enterprise AI adoption | https://www.datarobot.com | DetailsDataRobot is an enterprise automated machine learning (AutoML) platform that enables no-code/low-code machine learning workflows from data preprocessing to model building, tuning, and deployment. The main functions include automatic feature generation using "Feature Discovery," model ensemble construction, model interpretation using Explainable AI (SHAP values/LIME), and operational monitoring functions such as drift detection, and it automatically compares and optimizes up to several hundred types of algorithms. It can also be incorporated into organizations through the API and Python/R SDK, so not only data scientists but also business users can prototype and evaluate models on the GUI.
The companies that have introduced it are diverse, including finance, manufacturing, retail, and healthcare, and it is used for use cases such as credit scoring, anomaly detection, sales forecasting, demand forecasting, and maintenance prediction. DataRobot Intelligence Cloud provides a cloud-native deployment environment and runs scalably on any Azure/AWS/GCP platform. It is fully compatible with on-premise versions and hybrid environments, and maintains a security system that complies with SOC2/ISO27001.
The pricing system is a subscription model according to company size and scope of use, and can be expanded in stages from the Proof of Concept (PoC) phase to enterprise deployment. In case studies, there are reports that Company On the other hand, due to the initial learning cost due to the platform's multifunctionality and the limited flexibility of custom algorithms, some companies are facing challenges in developing data pipelines and improving AI literacy across the organization. In the future, it is expected that AutoML will further improve its accuracy and strengthen AI governance functions (bias detection and accountability). |
| 32 | Hugging Face Transformers | Hugging Face | America | Provision of NLP models, fine tuning support | Multilingual support | Free / paid plans available | Number of monthly users not disclosed | Widely used by AI researchers and developers | https://huggingface.co | DetailsHugging Face Transformers is an open source library and platform that makes it easy to work with various machine learning models, including natural language processing (NLP). It offers thousands of pre-trained models and tokenizers, including BERT, GPT, T5, RoBERTa, DistilBERT, etc., allowing you to implement advanced NLP tasks such as text classification, question answering, text generation, summarization, translation, Named Entity Recognition (NER), and text embedding generation with just a few lines of code. It supports both PyTorch and TensorFlow, making model fine-tuning and deployment easy.
Hugging Face Hub hosts custom models and trained checkpoints created by the community, allowing one-stop model sharing and reuse. By using the Inference API, you can achieve high-speed inference on the cloud and choose a plan according to your latency requirements. By linking with the Datasets library and Tokenizers library, it also supports efficient loading and preprocessing of large corpora, and faster subword segmentation.
In companies and research institutions, it is used for a variety of purposes, including automatic response chatbots in call centers, sentiment analysis of customer feedback, document summarization, automatic classification of legal documents, and meta-analysis of academic papers. It is fully equipped with a Python/R client, JavaScript SDK, and REST API, and has a proven track record in deploying models that support multimodal (image + text) and on-premises/edge devices. In the future, it is expected that the Optimum library will evolve to improve the memory efficiency of LLM and facilitate custom private fine-tuning of large-scale models. |
| 33 | Stable Diffusion | Stability AI | England | Image generation from text, open source | Multilingual support | free | Number of monthly users not disclosed | Popular with creators | https://stability.ai | DetailsStable Diffusion is an open source text-conditional image generation model led and developed by Stability AI. Adopts Latent Diffusion Model (LDM) architecture and can generate high resolution (1024 x 1024 or higher) images from text prompts. A variety of community front ends and APIs such as AUTOMATIC1111, Diffusers, and Stable Horde are in place, and self-hosted operations in on-premise GPU environments and clouds are becoming popular.
Academic papers and community resources introduce highly customized generation using class conditions, Inpainting, Image-to-Image, and ControlNet (pose/contour/depth information), and it is used in a wide range of use cases such as game asset generation, advertising banner creation, product design prototyping, art production, and automatic generation of educational visual materials. In terms of safety, progress is being made in developing NSFW content filtering and usage policies for licensed datasets, and higher quality models such as Stable Diffusion XL (SDXL) are now available for businesses.
The hurdles to implementation are GPU resource management and adjusting library dependencies, but we have prepared an environment that makes it easy to get started, such as deployment with Hugging Face Spaces and Lambda Labs' self-hosting service. For commercial use, it is necessary to establish legal risk management and product quality assurance processes, and future challenges include automatic optimization of text prompts (prompt generation) and standardization of safety control frameworks. |
| 34 | Qodo | Qodo | Israel | Code test case generation and development support using AI | English Center | Free trial available / Paid plan available | Number of monthly users not disclosed | Contributes to improving developer testing efficiency | https://www.qodo.ai | DetailsQodo is an AI-powered software test automation platform that consistently provides test case generation, test scenario optimization, and advanced analysis of execution results. It automatically generates test scripts from natural language and spreadsheet test requirements and is compatible with major test frameworks such as Selenium, Playwright, and Appium. It is easy to set up parallel test execution and integrate into CI/CD pipelines, and is reported to reduce testing effort by 40-60%.
The AI engine analyzes test results, clusters bug occurrence patterns, and visualizes gaps in test coverage. Additionally, it automatically summarizes failure reproduction steps and detects UI differences by comparing screenshots, improving the debugging efficiency of quality assurance teams. The dashboard allows real-time monitoring of test suite health indicators (pass rate, flake detection rate, test execution time), which helps identify quality bottlenecks.
The companies that have introduced it are diverse, including SaaS services, financial institutions, and game development companies, and it is especially useful in agile development environments where releases are frequently performed. The price is a subscription model depending on the number of tests executed, and it is compatible with both on-premises and cloud. In the future, we plan to enhance code change prediction using AI, automatic test refactoring functions, and natural language interfaces for end-to-end testing. |
| 35 | Figstack | Figstack | America | Code explanation, refactoring support | English Center | Free / paid plans available | Number of monthly users not disclosed | Helps developers understand code | https://figstack.com | DetailsFigstack is an AI pair programming tool that helps with code reading and refactoring. Connect GitHub and GitLab repositories to automatically review pull requests, analyze dependencies, and generate documentation for functions and classes. Especially for complex code bases, it extracts methods, recommends variable naming, detects redundant code, and suggests deletion, improving developer productivity.
AI combines abstract syntax tree (AST) analysis and large-scale language models to understand the intent of the code and generate suggestions. Display comments and correction suggestions within the IDE (VS Code plugin) in real time, enabling a quick review cycle. In case studies, it has been reported that Company Y reduced code review time by 50% by using Figstack, and Company Z reduced the bug recurrence rate after refactoring by 30%.
Pricing plans include a free community version and a paid plan for enterprises, with the paid plan offering on-premises deployment, private model training support, and secure SAML authentication. In the future, roadmap will include code conversion support between multiple languages and enhancements to automatic test code generation functions. |
| 36 | Codiga | Codiga | America | Code quality analysis, security check | Multilingual support | Free / Premium ($14.90/month) | Number of monthly users not disclosed | Contributes to improving the code quality of the development team | https://www.codiga.io | DetailsCodiga is a platform that automates code quality analysis and security checks. Integrated detection of code dead paths, potential security vulnerabilities (SQL injection, XSS, etc.), and best practice violations with hundreds of static analysis rules and a custom rules engine. Easy integration into your CI pipeline for real-time feedback from the pull request stage.
The dashboard visualizes code quality scores, security risk levels, and technical debt metrics to understand the health of each project at a glance. AI-based refactoring suggestions generate safe code rewrite patches and reduce manual effort for developers. Since it supports multiple languages (JavaScript, Python, Java, C#, Go, etc.), it can be applied even in large-scale microservice environments.
Companies using the service have reported results such as ``30% reduction in code review costs,'' ``20% increase in security detection rate,'' and ``15% reduction in post-release bug rate.'' Paid plans that offer dedicated model private cloud deployment and SLA guarantees are popular for enterprise implementation. In the future, it is expected that AI will improve the accuracy of code dependency analysis and strengthen security checks specific to front-end frameworks. |
| 37 | DeepCode AI | Snyk | America | Code review and bug detection using AI | Multilingual support | Free / paid plans available | Number of monthly users not disclosed | Supported by security-minded developers | https://www.deepcode.ai | DetailsDeepCode AI (now Snyk Code) is an AI-based code review service that performs mass analysis of open source code. Using our unique large-scale neural network model, we detect quality and security issues in function units with high accuracy and propose specific patches. Integrations with GitHub, GitLab, and Bitbucket automatically conduct reviews on pull requests.
DeepCode's strengths are training data based on hundreds of millions of lines of open source analysis and problem detection algorithms that take causal relationships into account. Detected vulnerabilities are assigned OWASP Top 10 category tags, allowing development teams to start with high-priority issues. The Enterprise plan supports on-premises environments and can be used with sensitive codebases.
Companies that have adopted it have reported results such as "vulnerability detection speed 5 times faster than conventional methods" and "security incident rate reduced by 40%." Future plans include automatic correction pull request generation using AI and expansion of continuous security monitoring functions for each CI/CD stage. |
| 38 | AskCodi | Assistiv.ai | Brazil | Code snippet generation and documentation creation support | English Center | Free / paid plans available | Number of monthly users not disclosed | Contributes to improving developer productivity | https://www.askcodi.com | DetailsAskCodi is an AI-powered development support platform that generates code snippets, test cases, and documentation from natural language prompts. The unique LLM-based AI engine understands the developer's project context and suggests appropriate code examples. It can be used with VS Code and Neovim plugins, and code completion and error correction assistance are also available in a chat format.
A feature of AskCodi is that it analyzes Git history and existing codebase and selects the optimal code path. It supports multiple languages and frameworks (JavaScript, Python, Java, Go, etc.) and automatically takes type safety and dependencies into account. It is offered in a freemium model, with premium plans including support for private repositories, custom model training, and team collaboration features. |
| 39 | CodeGeeX | THUDM | China | Multilingual code generation AI | Multilingual support | free | Number of monthly users not disclosed | Useful for multilingual developers | https://github.com/THUDM/CodeGeeX | DetailsCodeGeeX is a large-scale multilingual AI coding model developed by THUDM (a research team at Tsinghua University) that learns from tens of billions of lines of source code, including GitHub code. It is characterized by enhanced support for multilingual comments such as English, Japanese, and Chinese, and by extending VS Code, it realizes line-by-line and function-by-function completion, comment generation, code translation, and document summarization. It is open source and allows for model customization and self-hosting. |
| 40 | CodeT5 | Salesforce Research | America | Code generation, translation, summarization support | Multilingual support | free | Number of monthly users not disclosed | Used by AI researchers and developers | https://github.com/salesforce/CodeT5 | DetailsCodeT5 is a text/code compatible transformer model released by Salesforce Research that specializes in tasks such as source code completion, code repair, documentation generation, and code search. Pre-trained models are published on Hugging Face Hub, and datasets for fine-tuning are also provided. An architecture that combines NLP techniques and code parsing reduces inference costs while maintaining high accuracy. |
| 41 | Replit | Replit, Inc. | America | Browser-based development environment, equipped with AI auxiliary functions | Multilingual support | Free / paid plans available | Number of monthly users not disclosed | Popular with educational institutions and individual developers | https://replit.com | DetailsReplit is a browser-based integrated development environment (IDE) platform that covers everything from learning programming to prototyping, team development, and hosting. Supports multiple languages (Python, JavaScript, Ruby, Go, etc.) and offers thousands of template projects and package manager integrations. It integrates code completion, code generation, and bug fixing support functions using Replit AI (formerly Ghostwriter), and can generate code snippets even from natural language prompts. The multiplayer coding feature allows for real-time pair programming and workshops, and it is increasingly being adopted by educational institutions and hackathons. Furthermore, Replit Deploy allows you to publish web apps and APIs with one click, creating a workflow that allows you to quickly conduct market research and idea verification. In addition to the free plan, we offer student plans (Hacker, $7/month), Teams plans ($12/month/user), etc. to support team development and organizational use. It also supports collaboration with GitHub and Google Classroom, as well as OAuth login, and the number of implementation cases at educational institutions and companies is increasing. |
| 42 | OpenAI Codex | OpenAI | America | Code generation from natural language, GPT-3 based | English Center | Free (API usage is charged) | Number of monthly users not disclosed | Used to support coding for developers | https://openai.com/blog/openai-codex | DetailsOpenAI Codex is a large-scale language model based on GPT-3 with enhanced code understanding and generation capabilities, and can generate source code for Python, JavaScript, Go, Ruby, etc. from a natural language prompt. It can be called via API and is the basis of many coding support services such as GitHub Copilot and ChatGPT Code Interpreter. Codex outputs practical functions and scripts from program specifications, comments, and interactive instructions, and is suitable for a variety of DevOps use cases such as automatic generation of command line tools, test code creation, and bug fix pitch generation. Adopts a pay-as-you-go model based on usage with a token-based fee structure. On the security front, OpenAI's Use-Case Policy guidelines limit the handling of dangerous code and sensitive information. In the future, we plan to fine-tune Codex, provide private models, and deploy optimization to edge devices, which is expected to further deepen the AI-driven software development experience. |
| 43 | Sourcegraph Cody | Sourcegraph | America | Code search, navigation assistance | Multilingual support | Free / paid plans available | Number of monthly users not disclosed | Useful for managing large codebases | https://sourcegraph.com/cody | DetailsSourcegraph Cody is an AI coding assistant integrated with the code search platform Sourcegraph that searches for context from across your repositories in real time, delivering code snippets and documentation summaries in response to natural language prompts. Vectorize your codebase with the built-in Embeddings engine to speed up matching of related files and functions. It can be used with IDE plugins (VS Code, JetBrains products) and Web UI, and supports access management using API keys and SSO authentication. For teams, it includes security and compliance features to ensure collaboration on private repositories. Use cases include searching for usage examples of inexperienced APIs, automatic code reviews, bug fix recommendations, test coverage analysis, and refactoring proposal generation. Integration with Sourcegraph Enterprise helps improve knowledge sharing and productivity in large-scale monorepository operations. |
| 44 | Intellicode | Microsoft | America | AI code completion and style guide support | Multilingual support | free | | | | DetailsIntellicode is an AI-based code completion engine integrated into Microsoft Visual Studio / VS Code that learns your code patterns, shared repositories, and OSS datasets to provide advanced line-by-line and signature-level intellisense. The Team Completions feature builds a private model on the cloud that incorporates organization-specific coding styles and naming conventions, supporting company-wide deployment. It has a wide range of features, including contextual code suggestions, argument-named completion, documentation generation, and automatic code style enforcement. Included in the Visual Studio Professional/Enterprise license and available for free as an extension in VS Code. Works with version control and CI/CD pipelines to reduce review man-hours by analyzing differences during pull requests and automatically inserting warning comments. In the future, we plan to further integrate AI into the entire development experience, including strengthening collaboration with GitHub Copilot and adding AI debugger functionality. |
| 45 | Lumen5 | Lumen5 | Canada | Video creation platform using AI | English Center | Free / Premium ($19/month) | Number of monthly users not disclosed | Popular with marketers and content creators | https://www.lumen5.com | DetailsLumen5 is an online video creation platform that automatically converts text and blog posts into videos. AI extracts the main points of the manuscript and automatically selects the most suitable images, video clips, and BGM for each scene. Equipped with a wide variety of templates, fonts, and transitions, the brand kit function allows you to consistently reflect your logo and color palette. Fine adjustments can be made intuitively using the drag-and-drop UI, and automatic subtitle generation and text-to-speech (TTS) functions are also integrated. It is used for a wide range of purposes, including social media advertisements, educational content, and in-house training videos, and there are reports of cases where the initial man-hours for video production can be reduced by more than 90%. Pricing plans include Free, Starter ($19/month), Professional ($59/month), and custom plans for Enterprise, and higher-end plans that include API integration, SAML authentication, and dedicated support are increasingly being implemented by large companies. |
| 46 | QuillBot | QuillBot | America | AI-powered text rewriting, summarization, and grammar checking | English Center | Free / Premium ($9.95/month) | Number of monthly users not disclosed | Widely used by students and writers | https://quillbot.com | DetailsQuillBot is an AI writing tool whose main function is paraphrasing (paraphrasing) sentences. Restructures user-entered text in multiple stylistic modes (standard, formal, creative, etc.) to ensure originality and improve readability. It also integrates grammar checking, spelling correction, writing style suggestions, and vocabulary enhancement functions, and provides writing support in multiple languages, mainly English. The "Summarizer" function summarizes long texts and extracts important points. "Grammar Checker" corrects grammar and syntax errors in English sentences in real time. Compatible with Chrome extensions, Microsoft Word add-ins, and Google Docs integration for seamless integration into your writing flow. The free plan also allows you to paraphrase and summarize several times a day, while the premium plan ($95 per year) unlocks unlimited characters, batch processing, and plugin functionality. It is used by a wide range of users, including educational institutions, writers, and SEO marketers, and many say it has improved their work efficiency by 30%, but caution must be taken when using technical terms. |
| 47 | Looka | Looka | Canada | Logo design and brand kit creation using AI | English Center | Free trial available / Paid plan available | Number of monthly users not disclosed | Popular with startups and small businesses | https://looka.com | DetailsLooka is an AI-powered logo and branding design platform that allows users to automatically generate dozens to hundreds of logo ideas by simply entering the brand name, industry, and preferred colors and styles. Additionally, the generated logo can be instantly deployed on various templates such as business cards, social media images, website headers, etc. AI combines machine learning models and vector processing techniques to suggest balanced layouts based on industry trends and design principles. Logo adjustments can be made intuitively using the drag-and-drop UI, and fonts, colors, and icons can be freely customized. Once your design is complete, you can download it in a variety of file formats including SVG, PNG, and EPS, and a commercial license is included. With the free plan, you can only get low-resolution samples, but with the standard plan (starting from $20/month) and premium plan (starting from $65/month), you can get all the features and rights, and also supports team licenses. |
| 48 | Fathom | Fathom | America | Automatic meeting summary and note creation using AI | English Center | Free / paid plans available | Number of monthly users not disclosed | Supported by business professionals | https://fathom.video | DetailsFathom is an AI meeting assistant that works with online meeting platforms (Zoom, Teams, Google Meet) to record, transcribe, and automatically summarize meetings in one click. In parallel with real-time subtitle display, it automatically organizes "highlights", "action items", "key notes", etc. after the meeting ends, and generates a report for sharing. By using natural language processing (NLP) to classify topics and identify speakers, you can instantly extract only the necessary information from long meetings. Generated reports can be exported in PDF or Google Docs format, and integrates with numerous collaboration tools such as Slack and Notion. Even free users can record 30 minutes per month, and the Pro plan ($12/month) offers real-time summaries and unlimited meeting recordings. It received a 4.6/5 in G2 reviews, and is being used more frequently in customer success and remote teams, with people saying that their follow-up efficiency after meetings has improved by an average of 50%. |
| 49 | Nyota | Nyota | Unknown (under investigation) | AI-based note-taking and meeting assistant | English Center | Free trial available / Paid plan available | Number of monthly users not disclosed | Contributing to improving the efficiency of corporate meetings | https://nyota.ai | DetailsNyota is an AI-based personal research assistant tool that instantly searches and summarizes information on the web to support projects and report creation. Users can simply enter questions in natural language, analyze information from multiple sources, and get summaries with evidence. Perform deep web searches on academic papers, news articles, blogs, official documents, and more to automatically extract the data points you need. The source URL of the data is clearly indicated and a list of citations is automatically generated, making research work for paper writing and business reports much more efficient. Available as a browser extension or desktop app, team plans allow you to share and collaborate on research results. The free plan gives you 10 queries per day, while the Pro plan (starting at $15/month) unlocks unlimited queries, advanced filtering, and file upload analysis features. |
| 50 | Sudowrite | Sudowrite | America | Creative writing support using AI | English Center | $10/month? | Number of monthly users not disclosed | Popular with writers and screenwriters | https://sudowrite.com | DetailsSudowrite is an AI writing platform designed for writers and storytellers to help them with every phase of fiction writing, including plotting, characterization, text generation, rewriting, and adding dialogue. “Describe” generates scene descriptions, “Expand” makes short sentences more detailed, “Fix Grammar” proofreads grammar, and “Brainstorm” generates ideas to eliminate Writer’s block. Chrome extensions and Scrivener integration are also available, allowing you to utilize AI functions without leaving your writing environment. Pricing starts with a Starter plan of $10 per month, and a Pro plan ($25 per month) including unlimited tokens and team licenses are also available. Users have reported that their writing speed has doubled, and their word count has increased by an average of 30%, contributing to a qualitative improvement in the creative process. |
| 51 | Rytr | Rytr | India | Text generation and copywriting support using AI | Multilingual support | Free/Premium ($29/month) | Number of monthly users not disclosed | Advocated by marketers and writers | https://rytr.me | DetailsRytr is an AI-based writing assistant that can generate, paraphrase, and proofread a variety of text content such as blog articles, social media posts, ad copy, emails, product descriptions, and more. Users simply specify a prompt, tone, style, and length and get a high-quality draft within seconds. It offers 50+ languages and 30+ use case templates, and also integrates SEO optimization and grammar checking. In addition to being available as a Chrome extension, web app, and WordPress plugin, it also supports Zapier integration and automated workflow construction via API. Pricing starts with a free plan (generates 10,000 characters per month), and there are multiple plans such as Starter ($9/month), Growth ($29/month), Pro ($49/month), and high-priced plans that include unlimited characters and team collaboration features. We have received many comments from companies and freelancers who have adopted it, such as ``60% reduction in article creation time'' and ``stabilized writing quality.'' |
| 52 | Filmora | Wondershare | China | AI-powered video editing software | Multilingual support | Free trial available / Perpetual license $69.99? | Number of monthly users not disclosed | Widely used by everyone from beginners to professionals | https://filmora.wondershare.com | DetailsFilmora (Wondershare) is a video editing software for beginners and professional users, with AI features such as automatic scene detection, motion tracking, background remover, and automatic text summary caption generation. In addition to intuitive operation with a drag-and-drop UI, it is equipped with AI effects (AI portrait, style transfer, audio separation), and is compatible with a wide range of content from short videos to long-form content. It comes with over 500 transitions, effects, and templates, as well as specialized features such as multicam editing, color grading, audio mixing, and more. It supports cloud collaboration functions and team project sharing, and has been used in many educational institutions and marketing departments. You can choose from a perpetual license ($69.99/Yr) or an annual subscription plan, and we also offer annual licenses for businesses and plans for educational institutions. |
| 53 | OpusClip | OpusClip | America | Automatic generation of video clips using AI | English Center | Free / paid plans available | Number of monthly users not disclosed | Support for content creators | https://www.opus.pro | DetailsOpusClip is a tool that uses AI to analyze long videos, automatically extracts highlight scenes, and generates videos for short clips and reels. It uses a combination of camera movement, facial expressions, and audio analysis to detect important moments from the original uploaded video, and automatically inserts optimal transitions and text overlays. You can export vertical/horizontal formats for each SNS platform with one click, and it also has caption generation and hashtag suggestion functions. It operates as a web app and can be integrated into internal systems via API. Starting at $20 per month after a free trial, the Enterprise plan supports unlimited exports and custom branding. |
| 54 | Wordtune | AI21 Labs | Israel | Sentence rewriting and grammar improvement using AI | English Center | Free / Premium ($9.99/month) | Number of monthly users not disclosed | Popular with students and business people | https://www.wordtune.com | DetailsWordtune is a writing support tool developed by AI21 Labs that provides text rewriting, summarization, translation, and tone adjustment features. Users can change readability and tone while preserving the meaning of the text by simply highlighting a sentence and choosing a different phrasing from multiple options. Integrate with Chrome extensions, Microsoft Word, Google Docs, Outlook, and more to keep your writing flow flowing. The AI model has excellent language understanding ability and can be used for a variety of purposes such as switching between formal and casual, shortening sentence length, and emphasizing key points. The free plan has a monthly usage limit, while the premium plan ($24/month) offers unlimited rewrites, team sharing, and a management console for organizations. |
| 55 | Vista Social | Vista Social | America | AI-powered social media management platform | English Center | Free / paid plans available | Number of monthly users not disclosed | Advocated by marketers | https://vistasocial.com | DetailsVista Social is a social media management platform that optimizes content calendaring, scheduling, and analytics with AI. Helps maximize engagement of SNS posts with AI caption generator and hashtag suggestion function. Compatible with major SNS such as Instagram, Facebook, Twitter, LinkedIn, YouTube, etc., allowing cross-account management on one screen. It visualizes engagement rates, post performance, and follower increase predictions on a dashboard, and also has an automatic report creation function. We offer free plans, pro plans ($49/month), team plans ($99/month/5 users), etc., and API integration and white label reporting are available on higher-tier plans. |
| 56 | FeedHive | FeedHive | Denmark | Create and schedule social media posts using AI | English Center | Free trial available / Paid plan available | Number of monthly users not disclosed | Popular with social media managers | https://www.feedhive.io | DetailsFeedHive is an AI-powered social media content scheduler that unifies everything from post planning to automatic posting and analysis. One-click execution of AI captions, image selection, hashtag generation, and optimal posting timing suggestions. Compatible with Twitter, LinkedIn, Instagram, and Facebook, and tracks post performance, click-through rates, and conversions in real time. It is also equipped with a content calendar and team collaboration functions, and supports large-scale team operations with role-based access control. Pricing is Starter Plan ($29/month), Pro Plan ($99/month), Enterprise Plan (custom quote), and a free trial is also available. |
| 57 | Deep Research | Deep Research | America (estimated) | Research support tools using AI | English Center | Free trial available / Paid plan available | Number of monthly users not disclosed | Supported by researchers and analysts | | DetailsDeep Research is a research assistant that searches academic papers, news articles, patents, official documents, etc. on the web, and uses AI to summarize and analyze them. Enter keywords or natural language queries to extract key points from related documents and automatically generate research reports with citations. PDF upload analysis function allows you to import non-public documents, and it has team sharing and annotation functions. The free plan unlocks 5 reports per day, and the pro plan ($39/month) unlocks unlimited queries, custom filtering, and API access. |
| 58 | Clay | Clay | America | Sales lead generation and management using AI | English Center | Free trial available / Paid plan available | Number of monthly users not disclosed | Popular with sales teams | https://www.clay.com | DetailsClay is a platform that extends contact management and CRM functionality with AI, automatically collecting and integrating email, SNS interactions, and public information on the web to generate a timeline for each contact. AI helps sales teams become more efficient by setting reminders, recommending next contact, and automatically generating conversation summaries. Integrates with Notion, Slack, Google Workspace, Salesforce, and more to provide a single source of truth. Free plan (100 contacts), Pro plan ($29/month) and above include unlimited contacts, team collaboration, and SAML authentication. |
| 59 | Bubble | Bubble | America | No-code app development platform | English Center | Free / paid plans available | Number of monthly users not disclosed | Support for startups and small businesses | https://bubble.io | DetailsBubble is a no-code web app building platform that provides drag-and-drop UI, workflow automation, and database management features. Recently, an AI plugin has been added, making it possible to automatically generate UI components and data processing logic code from natural language prompts. Standard support for third-party API integration, user authentication, and payment processing, as well as cloud hosting and scalability. Pricing is free, personal ($29/month), professional ($129/month), and team ($529/user/month). |
| 60 | Lovable | Lovable | America (estimated) | App design support tool using AI | English Center | Free trial available / Paid plan available | Number of monthly users not disclosed | Popular with designers | https://www.lovable.app | DetailsLovable is an AI app builder that automatically generates React/TypeScript code from text prompts and provides instant feedback with live previews. Equipped with unlimited revisions, GitHub/Supabase integration, and edge deployment features, it allows for consistent front-end and back-end development. It runs on a browser, can be operated from templates, visual editors, and CLI, and facilitates team development with collaboration support functions. It offers free, standard ($19/month), and pro ($39/month) plans, with custom AI models and private repositories available on the pro plan. A sample project by the DevRel team is available on GitHub, and you can try out all the features with a 14-day free trial. |
| 61 | Cursor | Cursor | America | AI code completion and error checking | English Center | Free / paid plans available | Number of monthly users not disclosed | Support for developers | https://www.cursor.so | DetailsCursor is an AI-powered next-generation IDE platform that provides code navigation, refactoring, test generation, debugging assistance, and more. It enables code completion and function generation from natural language prompts, and integrates with GitHub repositories to accelerate development while maintaining consistency across your codebase. It has particularly strong cross-repository assistance in large-scale monorepo environments, and supports multiple languages (Python, JavaScript, TypeScript, Go, etc.). The AI engine uses a hybrid approach of code AST and LLM to improve the accuracy of context understanding. Integration into CI/CD pipelines also enables automated review of pull requests and expanded test coverage. Pricing is based on a freemium model, and basic code completion and refactoring suggestions are available even in the free version. The Professional plan ($15/user/month) offers advanced features such as support for private repositories, team collaboration, and custom AI model training. Companies that have adopted it have reported an average reduction in code review time of 30% and an increase in bug discovery rates of 15%, and future enhancements are planned for AI pair programming and real-time collaboration. |
| 62 | v0 | v0 | America (estimated) | Application development platform utilizing AI | English Center | Free trial available / Paid plan available | Number of monthly users not disclosed | Popular with development teams | https://www.v0.dev | Detailsv0 (v0.dev) is a request-driven API development platform designed for developers. AI supports endpoint design and validation logic generation while automatically generating Swagger and OpenAPI specifications. Enter your API's schema definition from the CLI or web UI and instantly generate backend code snippets (Node.js, Python, Go, etc.), documentation, and test cases. The generated API can be deployed with one click on Serverless Framework, Express, FastAPI, etc., and also integrates authentication/authorization, logging, and automatic scaling functions. The AI engine analyzes the existing modules of the project and automatically generates the code for the linked database (SQL/NoSQL) and cache layer. Freemium lets you try out basic features, while the Pro plan ($19/month) offers custom schema templates, advanced security rule generation, and team management features. In many cases, companies that have adopted it have reduced API development man-hours by 50% compared to conventional methods, and the time required to prepare documents has been significantly shortened.In the future, it is planned to be optimized for GraphQL support and a low-latency inference environment. |
| 63 | Clockwise | Clockwise | America | Calendar management and meeting coordination using AI | English Center | Free / paid plans available | Number of monthly users not disclosed | Contribute to team time management | https://www.getclockwise.com | DetailsClockwise is an AI scheduler that works with Google Calendar to optimize collaboration team schedules. Maximize individual and team efficiency with buffer time, automatic prioritization, and focus time features. You can add schedules and make adjustments using a natural language interface, and you can also operate from Slack or Microsoft Teams. AI learns the work patterns and priorities of members and automatically adjusts the balance between meetings and focused work. Paid plans offer team-wide visibility dashboards, space optimization algorithms, and automatic time zone adjustments for smart scheduling that moves beyond traditional calendar approaches. Companies that have introduced it have reported effects such as a 20% reduction in meeting time and a 15% increase in focused work time, and it is expected that collaboration with AI-based minutes analysis will be strengthened in the future. |
| 64 | Tidio AI | Tidio | Poland | Customer service chatbot powered by AI | Multilingual support | Free / paid plans available | Number of monthly users not disclosed | Popular for customer support for small and medium-sized businesses | https://www.tidio.com | DetailsTidio AI is an integrated customer support chatbot and live chat platform that provides automated AI responses, knowledge base search, and sentiment analysis. Automatically categorize inquiries from chat widgets installed on your website and instantly generate answers based on FAQs, product information, technical support, etc. You can edit the knowledge base from the management screen by dragging and dropping, and the AI regularly learns inquiry data to improve response accuracy. Paid plans offer Omni-Channel support (integrated channels such as email, Facebook Messenger, and Instagram DM), custom workflow creation, SAML authentication, and HIPAA/GDPR compliance functions. It has received high praise from users, including ``90% reduction in initial response time'' and ``50% improvement in chat operator productivity,'' and future implementation of voice recognition integration and real-time translation functions is planned. |
| 65 | Hiver | Hiver | India | AI customer service tool for Gmail | English Center | Free trial available / Paid plan available | Number of monthly users not disclosed | Contributing to improving the efficiency of corporate email response | https://hiverhq.com | DetailsHiver is a help desk collaboration tool that runs on Gmail and provides features such as automatic ticket sorting, priority setting, and agent recommendation using AI. Easily manage shared mailboxes (support@, sales@), visualize conversation history, and automate agent workflows. AI analyzes the intent of email content, routes to the appropriate agent or team, and even supports automatic response draft generation. The reporting function provides dashboards such as average resolution time, agent performance, and customer satisfaction prediction, contributing to operational improvements. Pricing plans include Lite, Professional, and Advanced, with the Advanced plan offering SLA guarantee, SAML authentication, and custom reporting functions. Companies that have adopted it have reported results such as a 40% reduction in support ticket processing time and a 15% increase in CSAT scores, and future chat support functionality within Slack/Microsoft Teams is planned to be expanded. |
| 66 | Textio | Textio | America | Job advertisement optimization tool using AI | English Center | Free trial available / Paid plan available | Number of monthly users not disclosed | Supported by HR personnel | https://textio.com | DetailsTextio is a writing assist platform for companies that provides polite analysis of job postings, sales emails, product descriptions, marketing content, etc. and real-time improvement suggestions using AI. Combine statistical language models with custom vocabulary lists to help maintain inclusive language and brand tone. The UI visualizes the scores of "strength," "empathy," and "diversity" for each phrase, making it easy to plan A/B tests for candidate phrases. Works with HR/CRM systems such as Salesforce, Workday, and Greenhouse for seamless integration into existing workflows. Delivered as a subscription model, companies that have adopted it have seen a 50% reduction in content lead time and a 20% increase in job applications. In the future, we plan to enhance multilingual support and support open-ended questions using the chat interface. |
| 67 | CVViZ | CVViZ | India | Resume analysis and candidate matching using AI | English Center | Free trial available / Paid plan available | Number of monthly users not disclosed | Popular with recruiters | https://cvviz.com | DetailsCVViZ is a recruiting platform that automates the hiring process with AI, centralizing resume analysis, candidate matching, and interview scheduling. We use NLP and machine learning to analyze skill sets, experience, and cultural fit, and quantify the degree of match with job requirements. Data synchronization is easy with the two-way collaboration function with ATS (Applicant Tracking System), and automatic notification workflows via Slack and email can be built. The report function supports recruitment KPI management by creating a dashboard of recruitment pipeline status, candidate pool diversity indicators, interview pass rate, etc. Pricing starts with an entry plan ($99/month/10 jobs), and an enterprise plan for large businesses with custom APIs and on-premises deployment. Companies that have adopted the service have confirmed that it has reduced recruiting costs by 30% and shortened the hiring period by 20%, and in the future, AI interview analysis and chatbot interview functions are planned to be added. |
| 68 | Notion AI Q&A | Notion | America | Knowledge management and Q&A functions using AI | Multilingual support | Free / Available | | | | DetailsNotion AI Q&A is a feature that searches documents and databases within Notion in real time and automatically generates summaries and answers to questions in natural language. Just by typing "/ask" on a page, AI extracts relevant information and presents it in bulleted list format or body text format. AI leverages page metadata and tagging information to provide optimal answers based on context. Enterprise plans include integration with internal knowledge bases, custom dictionaries, and advanced access control settings. Organizations that have adopted it have reported results such as a 70% reduction in search time and improved accuracy in information searching, and in the future, multi-domain Q&A and natural dialogue interfaces are expected to be strengthened. |
| 69 | Otter.ai | AISense, Inc. | America | Real-time transcription and summarization of meetings | English, Japanese, Others | Free / Pro ($8.33~/month) / Business ($20~/month) | Many integrated companies | Zoom, Teams collaboration; many examples of implementation in educational institutions and companies | https://otter.ai | DetailsOtter.ai is an AI meeting assistant tool that transcribes audio from meetings and interviews in real time and automatically generates summaries and action items. It seamlessly works with video conferencing services such as Zoom, Microsoft Teams, and Google Meet, providing comprehensive support from note-taking during the meeting to follow-up afterwards. Adopting a Transformer-based speech recognition model, it also enables noise reduction and speaker identification, enabling accurate transcription of multi-person meetings. It has a wide range of functions to maximize the use of minutes, such as extracting important keywords from minutes, creating action checklists, and automatically generating summary sections. A variety of plans are available, from the free version to the enterprise version, and it also provides SAML SSO, ISO27001-compliant security operations, and GDPR compliance. It has a rich track record of implementation at educational institutions and major companies, and has been praised for its high productivity improvement effect. [Reference] https://otter.ai/blog/the-leading-ai-meeting-assistant-otter-ai-unveils-game-changing-productivity-boost-62-of-professionals-say-that-ai-saves-them-over-an-entire-month-of-work-each-year |
| 70 | Copy.ai | Copy.ai, Inc. | America | Short marketing content generation | Multilingual support | Free (7 days) / Pro ($36/month) | Employed by small and medium-sized companies | Shopify, Zapier collaboration | https://www.copy.ai | DetailsCopy.ai is a tool that automatically generates short content such as advertisements, emails, and social media posts, and includes dozens of templates and prompt libraries. The UI is intuitive, and customization such as changing the tone and setting prohibited words after generation is easy. Even with the free plan, you can generate 5000 characters per month. The Pro plan offers unlimited generation and team management, with over 1 million monthly global visits. [Reference] https://www.similarweb.com/website/copy.ai/ citeturn0search2 |
| 71 | Descript | Descript, Inc. | America | Audio/video editing and transcription | English | Free / Creator ($12/month) / Pro ($24/month) | Popular among creators | Vox Media, TED introduction | https://www.descript.com | DetailsDescript is the only platform that lets you edit audio and video using text. Generate a transcript using the automatic transcription function, and complete text editing to correct typos and cut out unnecessary parts. The Overdub function realizes speech synthesis using AI, allowing you to replace words without additional recording. Video editing links the timeline view and text view, allowing you to intuitively operate cuts, fades, and insert captions. [Reference] https://www.similarweb.com/website/descript.com/ citeturn0search4 |
| 72 | Replika | Luka, Inc. | America | AI chatbot/mental support | English, other | Free / Pro ($7.99/month) | Millions of users | Recognized as a mental health app | https://replika.ai | DetailsReplika is a chatbot that converses with users as their “personal AI friend.” Learning about emotions through dialogue and providing self-development and mental health support. It supports not only text chat but also voice chat and video chat, enhancing sentiment analysis and empathetic responses. The Pro Plan provides thematic topic guides and audio coaching functions, which are used for daily self-care and stress management. [Reference] https://www.similarweb.com/website/replika.ai/ citeturn0search5 |
| 73 | Murf.ai | Murf AI Inc. | India | AI speech synthesis platform | Over 110 languages/voice types | Free / Basic ($19/month) / Pro ($39/month) | Active in narration production | Available on Udemy, Coursera | https://murf.ai | DetailsMurf.ai is a platform that uses AI to synthesize voices for narration and podcasts. Supports over 110 voice types and multiple languages, and naturally reproduces pronunciation and intonation. By inputting an audio script, high-quality narration audio is generated in minutes and can be downloaded as an audio file. [Reference] https://www.similarweb.com/website/murf.ai/ citeturn0search18 |
| 74 | Typeform AI | Typeform S.L. | Spain | AI support for creating questionnaire forms | Multilingual support | Free / Essentials（$25/月） / Professional | Used for event management/research | HubSpot, Slack integration | https://www.typeform.com | DetailsTypeform AI provides the ability to automatically generate questionnaires and questionnaire forms using AI. Automatically create question sentences, suggest answer candidates, and set logical branching with one click. Form designs can be edited with drag and drop, and there are plenty of integrations with CRM and marketing tools. [Reference] https://www.similarweb.com/website/typeform.com/ citeturn0search9 |
| 75 | Qwen 2.5-1M | Alibaba Cloud | China | A high-performance LLM developed by Alibaba. Features multilingual support and highly accurate natural language processing. | | | Number of visits per month: unknown | Large-Scale Language Model (LLM) | https://qwen.alibaba.com | DetailsQwen 2.5-1M is a large-scale language model developed by Alibaba Cloud, with multilingual support and high-precision natural language processing capabilities. In particular, it demonstrates high performance in business applications and customer support, contributing to companies' operational efficiency. |
| 76 | Doubao 1.5 Pro | ByteDance | China | Chat AI
High-performance chat AI provided by ByteDance. Characterized by natural interaction and multifunctionality. | | | Number of visits per month: unknown | | https://doubao.bytedance.com | DetailsDoubao 1.5 Pro is a chat AI developed by ByteDance that combines natural interaction ability and versatility. It enables smooth communication with users and is used in a variety of applications, including customer support and education. |
| 77 | Kimi Chat | Moonshot AI | China | Chat AI
Chat AI developed by Moonshot AI. Accurately understand user intent and provide appropriate responses. | | | Number of visits per month: unknown | | https://kimi.moonshot.cn | DetailsKimi Chat is a chat AI developed by Moonshot AI that focuses on accurately understanding user intent and providing appropriate responses. It is increasingly being used in education and business settings, contributing to improving the user experience. |
| 78 | Yuanbao | Tencent | China | Chat AI
Chat AI provided by Tencent. It is widely used by users due to its integration with WeChat. | | | Number of visits per month: unknown | | https://yuanbao.tencent.com | DetailsYuanbao is a chat AI developed by Tencent that has been integrated with WeChat and is used by a wide range of users. It is used for a wide variety of purposes, from providing daily information to business support, and is becoming increasingly popular in China. |
| 79 | Manus | Monica | China | autonomous AI agent
An autonomous AI agent developed by Monica. Ability to perform complex tasks automatically. | | | Number of visits per month: unknown | | https://manus.monica.cn | DetailsManus is an autonomous AI agent developed by Monica that has the ability to perform complex tasks automatically. It contributes to the automation and efficiency of business processes and improves corporate productivity. |
| 80 | Ernie Bot 4.5 | Baidu | China | Large-Scale Language Model (LLM)
LLM developed by Baidu. Features high-precision natural language processing and multilingual support. | Chinese (simplified/traditional), English | Free plan, API usage is pay-as-you-go | Number of monthly visits: Approximately 2 million | Obtained Chinese government certification and has been introduced by many major domestic companies. | https://ernie.baidu.com | DetailsErnie Bot 4.5 is the most advanced large-scale language model (LLM) developed by Baidu, with extremely high accuracy in both natural language understanding and generation. This model supports both Chinese and English, and can perform a wide variety of tasks such as providing information, summarizing, translating, and writing support in response to various user inquiries. Through pre-learning using large-scale datasets, it covers specialized domain knowledge, and is increasingly being used as industry-specific solutions in fields such as finance, medicine, and legal affairs.
Provision via API uses a pay-as-you-go system based on usage, allowing developers to flexibly select the computing resources they need. Furthermore, with security and privacy protection in mind, encrypted data storage and access controls have been implemented. By leveraging Baidu's cloud infrastructure, we ensure scalability and high availability, and can handle large-scale traffic.
According to reports from companies that have implemented Ernie Bot 4.5, incorporating Ernie Bot 4.5 into customer support automated response systems has improved response quality by more than 20% and reduced the average inquiry response time by 30%. It is also increasingly being used as a multilingual chatbot and internal document search tool, contributing to improved productivity in companies. |
| 81 | KIBIT | FRONTEO | Japan | Natural language processing AI | Japanese, English | Inquiry required | Imported companies: more than 200 companies | Introduced by major companies such as the University of Tokyo and Nomura Holdings | https://www.fronteo.com | DetailsKIBIT is an advanced supervised machine learning platform developed by FRONTEO that aims to mimic and automate the decision-making process of experts. It is especially being used in fields that require specialized knowledge, such as legal document review, insurance/financial risk analysis, and classification of clinical trial data.
This system learns using past expert judgment data to suggest similar cases and detect anomalies. Extracts important keywords and patterns from documents and highlights high-risk areas. Streamline the review process and reduce human error. We also focus on UI/UX, allowing reviewers to see results through intuitive operations, and continuously improve the model by adding feedback.
Enterprise plans allow multiple users to manage permissions, share custom dictionaries, and tune for industry-specific terminology and regulatory requirements. In addition to being provided on FRONTEO's cloud service, it is also compatible with on-premises environments and can be operated in accordance with internal security policies.
Many financial institutions and law firms have implemented it, and there are reports of cases in which the man-hours required for review operations have been reduced by up to 50%. In addition, we provide expert support and model adjustment services during implementation, allowing you to achieve ROI in a short period of time. |
| 82 | Frase | Frase.io | America | Content optimization and SEO support platform. AI enables topic research, content brief generation, and competitive analysis. | Over 20 languages including English, Japanese, Spanish, etc. | Free plan, starting at $19.99/month | Number of monthly visits: Approximately 1.2 million | Can be linked with HubSpot and SEMrush | https://www.frase.io | DetailsFrase is an AI-based content optimization platform from Frase.io, designed for SEOs and content marketers. Automatically performs topic research based on keywords and themes entered by users, analyzes top search pages, and extracts important subtopics and questions. Based on the results, a brief for article structure is generated to help authors create comprehensive and effective content. |
| 83 | CoeFont ACCENT | HOYA Corporation | Japan | High-quality speech synthesis service specializing in Japanese. Provides a variety of voice qualities. | Japanese | Free plan, API usage is pay-as-you-go | Imported companies: more than 150 companies | Adopted by local governments and educational institutions | https://coefont.cloud | DetailsLumen5 is an AI platform that automatically converts text-based content into short-form videos.It takes the main points of articles, newsletters, and blog posts and generates social media videos of between 15 and 60 seconds. When users enter summary text on the platform, AI uses natural language processing technology to extract key phrases and automatically select the most appropriate images, video clips, and icons for each scene. In addition, text captions for background music and narration are automatically inserted to attract the attention of viewers. The intuitive drag-and-drop interface does not require specialized video editing skills and can be easily used by marketers and small businesses. Our extensive template library includes hundreds of designs for each industry and purpose, and you can also upload your brand colors, fonts, and logos. The AI-powered scene segmentation algorithm analyzes the context of the text to create visually appropriate cuts and automatically generate a video structure that tells a story. Team Plan is equipped with project sharing functions, commenting functions, and asset management functions to streamline collaborative production and review work among multiple members. The Enterprise plan provides security features such as SAML authentication and IP restrictions, dedicated support, and custom template creation services, making it suitable for large organizations. It can also be integrated into workflows via API, and can be integrated with CMS and marketing automation tools to automate dynamic content generation. According to official announcements, many companies have reported a 20% increase in views and a 15% improvement in click-through rates for social media campaigns, while reducing video production time by an average of 70%. As major implementation examples, HubSpot and Forbes used Lumen5 to mass produce newsletter summary videos and sales presentation videos, contributing to increased engagement with their own media. |
| 84 | Haptik | Haptik Inc. | India | AI chatbot platform. Specializing in customer support automation. | Over 15 languages including English and Hindi | Inquiry required | Imported companies: more than 500 companies | Adopted by major telecommunications companies and EC sites | https://www.haptik.ai | DetailsHaptik is a conversational AI platform for enterprises originating from India, and its biggest feature is that it can easily automate customer support and develop chatbots. Equipped with an NLU (natural language understanding) engine, it is multilingual (over 15 languages including English and Hindi) and supports multiple channels including phone, web chat, WhatsApp, and Facebook Messenger. The drag-and-drop visual flow editor allows even non-engineers to design scenarios, and a variety of templates are available, including automatic FAQ responses, ticketing, and upsell recommendations. The analysis dashboard supports continuous tuning cycles by visualizing user intent detection accuracy, resolution rate, exit points, etc. in real time. The Enterprise Plan comes with an SLA guarantee and a dedicated customer success manager to assist with implementation, and has been adopted by over 500 companies, including major telecommunications companies and e-commerce sites.
Reference site: https://www.haptik.ai, https://ja.wikipedia.org/wiki/Haptik |
| 85 | iFLYTEK | iFLYTEK Co., Ltd. | China | Speech recognition and synthesis platform. Multilingual voice AI. | Chinese (multiple dialects), English, etc. | Free plan, API usage is pay-as-you-go | Number of monthly visits: Approximately 1 million | Adopted by educational institutions and call centers | https://www.iflytek.com | DetailsiFLYTEK is a leading voice AI company based in China, and is a platform that comprehensively provides voice recognition, speech synthesis, translation, dialogue engines, etc. In particular, our highly accurate speech recognition technology that supports multiple dialects of Chinese (Mandarin, Cantonese, Shanghainese, etc.) has received high praise worldwide. It can be embedded through SDKs and APIs, and is increasingly being used in various environments such as mobile apps, smart speakers, and call center systems. Synthetic speech achieves rich expressiveness and natural intonation, and covers a wide range of use cases such as the education market, traffic information, and automatic response IVR. In addition, real-time voice transcription and translation functions are used on online learning platforms and in medical settings to reduce communication barriers. In addition to a free plan, we offer a pay-as-you-go API, and we receive approximately 1 million visits per month, with many implementations at educational institutions and major call centers.
Reference site: https://www.iflytek.com, https://ja.wikipedia.org/wiki/KadaiTanhi |
| 86 | Voiceful.io | Voiceful, Ltd. | Israel | AI voice synthesis for music and narration. Audio generation for creatives. | English | Free plan, starting from $20/month | Number of monthly visits: Approximately 200,000 | Used by musicians and podcasters | https://voiceful.io | DetailsVoiceful.io is an AI voice platform for creatives from Israel, specializing in voice synthesis for music production, narration, and podcasts. In addition to being able to generate emotionally expressive narration audio from text, it also has the ability to automatically generate melodies and harmonies, making it highly praised by musicians and music producers. A web-based editor lets you fine-tune your voice style and speed, and SSML (Speech Synthesis Markup Language) support lets you control pronunciation and pauses. It is easy to incorporate into game audio and AR/VR content through the API, and commercial licenses are available with free plans and paid plans starting from $20 per month. Case studies include podcast production companies and animation production studios, and there are reports of cases where production costs have been reduced by more than 50% through high-quality audio expression.
Reference sites: https://voiceful.io, https://www.producthunt.com/posts/voiceful-io |
| 87 | Uizard | Uizard Technologies | Denmark | An automatic UI design generation tool using AI. Prototype creation from wireframes. | English | Free plan, starting from $12/month | Number of monthly visits: Approximately 400,000 | Compatible with Figma and Sketch | https://uizard.io | DetailsUizard is an AI-based automatic UI design generation platform developed by Denmark's Uizard Technologies. Users upload hand-drawn wireframes or screenshots, and AI transforms them into modern interfaces and automatically builds them into component-by-component, reusable design systems. This allows designers and non-designers to quickly create prototypes and export them to Figma or Sketch formats.
It also features an intuitive UI editor with drag-and-drop operations to centrally manage color palettes, font styles, and icon sets. Utilize team collaboration and commenting functions to streamline design reviews with multiple members. Plus, with Slack integration and version control features, you can track feedback history and easily roll back design changes.
Uizard's AI engine continuously learns to adapt to brand guidelines for each project, making automatically generated results more customizable. Companies that have adopted it have reported that it reduces prototype creation time by an average of 65% during the early stages of web and mobile app development. It is also highly praised for automating the workflow from design to code generation through API integration and facilitating collaboration with engineers. |
| 88 | Hotpot.ai | Hotpot AI | Hongkong | AI image generation/editing tool. Art, photo retouching, text to image. | English | Free plan, starting at $9.99/month | Number of monthly visits: Approximately 300,000 | Chrome extension available | https://hotpot.ai | DetailsHotpot.ai is an AI-based image generation and editing toolset provided by Hotpot AI from Hong Kong. In addition to being able to generate illustrations and photo-like images from text prompts, it also has a variety of functions such as background removal, color correction, and cartoonization of existing images. Users can operate intuitively on the browser, and color adjustments and retouching can be completed with one click.
In particular, the artwork creation function uses GAN (Generation Opposed Network) to provide detailed and creative visuals. Batch processing mode allows you to convert hundreds of images at once, making it suitable for optimizing product images for e-commerce sites and creating materials for social media posts. In addition to being able to edit images directly on web pages via a Chrome extension, it is also possible to embed them using the API.
The corporate plan includes a team account management function to manage asset sharing and comment history for each project. There have been multiple reports of cases where the introduction of Hotpot.ai reduced design man-hours by an average of 50% and shortened the time to launch. |
| 89 | BeyondWords | BeyondWords Ltd. | England | AI voice synthesis and narration platform. High quality TTS. | English | Free plan, starting from $19/month | Number of monthly visits: Approximately 250,000 | SSML compatible | https://beyondwords.io | DetailsBeyondWords is an AI speech synthesis platform developed by BeyondWords Ltd. in the UK.It provides advanced voice control compatible with SSML, with a focus on the TTS (Text-to-Speech) function that converts text into natural narration. Users can select their favorite voice from a rich voice library and make detailed adjustments such as speed, pitch, and pause.
In addition to inputting text directly on the platform, it also has the ability to upload Markdown format documents and generate them in bulk. API integration makes it easy to integrate into news sites, e-learning platforms, and podcast hosting services, and automates the audio conversion of dynamic content. Team plans allow you to collaborate on audio projects and set access permissions.
BeyondWords' TTS engine uses a deep learning-based acoustic model and supports stress intonation and emotional expression. Companies that have adopted it have reported that they have been able to reduce narration production costs by 70% and shorten release cycles. |
| 90 | Mistral | Mistral AI | France | large language model
API/Open source | | Free / paid plans available | | | https://mistral.ai | DetailsMistral is an AI startup from France, and although it was only founded in 2023, it is rapidly gaining attention. In particular, it is highly praised for its large-scale language model (LLM), which is completely open source and can be used commercially. Founded by engineers from Meta and Google, the company has become a symbol of innovation in the European tech scene. In 2024, we released a powerful language model called Mistral 7B, which was welcomed by many developers. Adoption cases in Europe and abroad are increasing, and it has the potential to become a company comparable to Google and OpenAI in the future. |
| 91 | AI21 Labs | AI21 Labs | Israel | Large-scale language model, sentence generation | | Free / paid plans available | | | https://www.ai21.com | DetailsAI21 Labs is an AI company based in Tel Aviv, Israel that develops large-scale language models specifically for natural language processing (NLP). A representative product is an LLM called Jurassic-2, which has performance comparable to OpenAI's GPT-3. AI21 provides highly accurate text generation and proofreading support tools, and is used in a wide variety of fields such as education, business, and writing. Another popular text rewriting tool is Wordtune, which is extremely useful for writers and marketers. As a startup from Israel, we are increasing our global presence. |
| 92 | Cohere | Cohere | Canada | natural language processing | | Free / paid plans available | | | https://cohere.com | DetailsCohere is an AI company based in Toronto, Canada that provides natural language processing technology as an API. In particular, its application to text generation, semantic search, and classification tasks using its unique LLM is progressing, and its simple interface optimized for developers is popular. The founder is from Google Brain and provides solutions for businesses with security and privacy in mind. It also has strengths in multilingual support and scalability, and is being introduced in a variety of industries such as finance, education, and medical care. |
| 93 | You.com | You.com | Switzerland | Search engine, AI assistant | | Free / paid plans available | | | https://you.com | DetailsYou.com is an AI search engine from Switzerland, a platform that integrates AI summarization, code generation, chat functions, and more. Designed with user privacy as a top priority, it aims to provide a highly accurate search experience while eliminating ads. AI chat (YouChat) is strong in question answering and creative writing, and also offers a code completion tool called YouCode for developers. It has been praised for leveraging Switzerland's neutral position and ethics to provide a search experience that stands out from Google. |
| 94 | H2O.ai | H2O.ai Inc. | America | Open source AutoML platform. Supports construction and deployment of machine learning models. | | | | | https://www.h2o.ai | DetailsOverview
H2O.ai is a company that provides H2O, an open source platform for machine learning and AI applications, and Driverless AI, helping automate decision-making with AI. It is designed to help companies and data scientists quickly build, deploy, and operate predictive models.
Main Products
H2O (Open Source): A multi-platform machine learning library that can be integrated with R, Python, Spark, etc. Supports classification, regression, clustering, anomaly detection, etc.
Driverless AI (commercial): Automated machine learning (AutoML) tool with automated feature engineering, hyperparameter tuning, model interpretability, and more.
H2O Wave: A web application development framework that enables rapid development of AI applications with no-code/low-code.
Use Cases
Financial Industry: Credit Scoring, Fraud Detection
Healthcare Industry: Disease Prediction, Medical Imaging
Manufacturing Industry: Demand Forecasting, Equipment Anomaly Detection
Retail Industry: Customer Segmentation, Inventory Optimization
Company Information
Founded: 2012
Headquarters: Mountain View, California, USA
Customer Examples: PayPal, Capital One, AT&T, Walmart etc. |
| 95 | IBM Watson | IBM | America | An AI platform that provides natural language processing, machine learning, and data analysis. | | | | | https://www.ibm.com/watson | DetailsOverview
IBM Watson is the collective term for artificial intelligence services and platforms provided by IBM, with a wide range of capabilities including natural language processing (NLP), speech recognition, machine learning, visual recognition, and more. It is used in many industries including healthcare, finance, manufacturing, education, and customer support.
Main Product Lines
Watson Assistant: Platform for building chatbots and virtual assistants. Used for voice dialogue and FAQ responses.
Watson Discovery: Automate the search and analysis of structured and unstructured data. Extract insights from business documents and reports.
Watson Natural Language Understanding (NLU): Provides NLP functions such as sentiment analysis, keyword extraction, and document classification.
Watson Studio: An integrated development environment that supports the design, training, and operation of machine learning models for data scientists.
Use Cases
Medical: Cancer diagnosis support (Watson for Oncology), electronic medical record analysis
Legal: Contract analysis and risk detection (Watson Legal)
Customer response: Optimizing customer service with AI chatbots
Education: Building personalized learning support tools
Company information
Founded: 1911 (Watson launched in 2011)
Headquarters: Armonk, New York, USA
Origin of name: From IBM founder Thomas J. Watson |
| 96 | Numerous | Numerous AI Inc. | America (estimated) | A spreadsheet AI tool that supports AI utilization in Google Sheets and Excel. | | | | | https://numerous.ai | DetailsOverview: Numerous is a tool that integrates AI functionality into spreadsheets (Google Sheets and Excel). Leverage the power of ChatGPT to automate tasks such as data classification, extraction, text generation, and data cleaning. This allows you to streamline repetitive tasks and save significant time.
Main features:
Data classification and extraction: Automatically extract and classify specific information from large amounts of data.
Text generation: Generates text such as email text or reports based on the contents of the specified cell.
Data Cleaning: Shape irregular data and convert it into a uniform format.
Marketing support: We support you in generating advertising copy and creating SEO content.
Use example:
Marketing team: Automatic generation of keywords and copy for advertising campaigns.
Sales team: Organize customer data and create analytical reports.
Human resources department: Classification of applicant data and automatic creation of feedback. |
| 97 | Zoom IQ | Zoom Video Communications, Inc. | America | AI features that help make meetings more efficient, such as summarizing Zoom meetings and providing insights. | | | | | https://explore.zoom.us/en/products/zoom-iq/ | DetailsOverview: Zoom IQ is an AI assistant feature provided by Zoom. Streamline your work during and after meetings by summarizing meetings, composing chat messages, suggesting tasks, and more. By collaborating with OpenAI, we have achieved advanced natural language processing.
Key features:
Meeting summary: Summarize meeting content in real time and share it with participants.
Create chat messages: Automatically generate follow-up messages based on meeting content.
Task proposal: Extracts action items discussed during the meeting and proposes them as tasks.
Example of use:
Project manager: Quickly grasp the main points of the meeting and use it for task management.
Sales team: Summarize the content of customer meetings and clarify next actions.
Human resources department: Summarize the content of the recruitment interview and use it to evaluate candidates. |
| 98 | tl;dv | tl;dv GmbH | Germany | An AI tool that automatically records and summarizes Zoom and Google Meet meetings. | | | | | https://www.tldv.io | DetailsOverview: tl;dv is a tool that records online meetings such as Zoom, Microsoft Teams, Google Meet, etc. and automatically creates minutes using AI. Supports over 30 languages and streamlines post-meeting tasks such as linking with CRM and drafting follow-up emails.
Main features:
Automatic minutes creation: Automatically generates minutes summarizing key points based on the recording of the meeting.
Multilingual support: Minutes can be created in over 30 languages.
CRM integration: Links with CRMs such as Salesforce and HubSpot to automatically update meeting information.
Follow-up email creation: Automatically generate a draft follow-up email based on the meeting content.
Use example:
Sales team: Automatically record the content of meetings with customers and clarify next actions.
Product team: Organize the contents of user interviews and use them to improve the product.
Human resources department: Records the content of recruitment interviews and uses them to evaluate candidates. |
| 99 | ELYZA Brain | ELYZA株式会社 | Japan | A text generation and summarization support tool that utilizes a large-scale language model (LLM) specialized in Japanese. | | | | | https://elyza.ai | DetailsOverview: ELYZA Brain is a Japanese-specific large-scale language model (LLM) developed by ELYZA in Japan. It excels in Japanese natural language processing and can handle a variety of tasks such as summarizing, generating, and classifying sentences. Customization for companies is also possible, contributing to increased business efficiency.
Main features:
Text summary: Summarize long sentences and extract the main points.
Sentence generation: Generates natural Japanese sentences based on specified themes and keywords.
Text classification: Analyzes the content of sentences and classifies them into categories.
Customization support: Models can be customized to meet the needs of your company.
Use examples:
Media companies: Summarizing and automatically generating news articles.
Customer support: Classification of inquiry content and automatic generation of correspondence.
Educational institutions: Automatic creation of teaching materials and evaluation of student reports. |
| 100 | AI GIJIROKU | 株式会社オルツ | Japan | An AI minutes creation tool that automatically transcribes and summarizes meetings. | | | | | https://gijiroku.ai | DetailsOverview: AI GIJIROKU is an automatic meeting minutes creation tool provided by alt Inc. in Japan. Record online meetings such as Zoom, Skype, Teams, etc., and AI will transcribe and translate them in real time. A smartphone app is also available so you can use it on the go.
Main features:
Real-time transcription: Converts conference audio into text in real-time.
Multilingual translation: multiple transcribed contents |
| 101 | ChatGPT for LINE | 株式会社ソフトバンク | Japan | A chatbot service that allows you to use ChatGPT on LINE. | | | | | https://chatgpt-line.me | DetailsDeveloper company/country
SoftBank Corp. (Japan)
Based on OpenAI's ChatGPT API, unique implementation to run on the LINE platform
Overview
ChatGPT for LINE is a bot service provided by SoftBank that allows you to easily use ChatGPT within the LINE app. Users can start a conversation using a GPT-4-based natural language generation model by simply entering "@ChatGPT" on the LINE talk screen. Incorporating unique conversation UI customization and collaboration functions with various APIs owned by Softbank (weather information, transit information, translation, etc.), it provides a one-stop solution for a wide range of scenarios from general chat functions to business support and daily life information acquisition.
Main features
Natural language dialogue: Realizes high-precision Japanese responses equivalent to GPT-4. Suitable for everything from everyday conversation to business document creation.
FAQ automatic response: Can be used as a customer support chatbot by linking with a FAQ database pre-registered by companies.
Multimedia search: Call external APIs such as weather, news, and maps from LINE's rich menu and return the results within the chat.
Business template: Equipped with a template function that automatically generates email drafts, quotation text, and drafts of meeting minutes.
Webhook collaboration: By linking with internal systems and CRM, it is possible to create a chat → internal data reference/update flow.
Technical background
API linkage layer: Place a proxy server for ChatGPT on the LINE Messaging API. We independently developed the necessary pre-processing and post-processing (context management, action determination).
Conversation context management: Stores long text history up to 4,000 tokens. The memory function allows past conversation history to be converted into sessions for each user ID.
Security: All communications are TLS encrypted. User data is held at a domestic data center and complies with the GDPR/Personal Information Protection Act.
Introduction Example
Major insurance company: Automatically responds to FAQ inquiries from policyholders 24 hours a day by linking with LINE official account. The response rate has increased by 35% compared to the previous model, and the operation cost has decreased by 20%.
Global manufacturing base: As a trilingual chatbot (English, Japanese, and Chinese), it is used for initial response at the internal help desk. Average initial response time reduced from 5 minutes to 1 minute.
Educational institutions: Adopted as a bot for entrance exam preparation and study consultation. Automatically generates explanatory texts and reference links according to question content to support students' self-study.
Challenges and future outlook
Preventing false information: It is essential to strengthen the fact-checking function of generated content. In the future, we are considering two-way synchronization with the knowledge base and integration of an expert-supervised API.
Optimization of conversation UX: Through further improvements to the talk design, we plan to implement more natural “complaints” and “expressive responses.”
Deepening business automation: Promote the "Chat RPA" concept that realizes business workflow completion from chat through RPA and OCR collaboration. |
| 102 | AI Picasso | 株式会社AI Picasso | Japan | A Japanese-compatible image generation AI tool that generates images from text. | | | | | https://ai-picasso.com | DetailsDeveloping company/country
AI Picasso Co., Ltd. (Japan)
An AI venture certified by the Ministry of Economy, Trade and Industry
Overview
AI Picasso is a creative AI tool from Japan that can automatically generate high-quality illustrations and stamps from text prompts. Users can obtain commercially usable PNG/SVG format materials in a few seconds by simply inputting ``a stamp of a cute Shiba Inu enjoying cherry blossoms under the cherry blossoms'' in Japanese. By combining GAN-based image generation technology with a uniquely trained custom model for Japanese creators, it faithfully reproduces "illustration-like touches" and "manga-like line drawings." Furthermore, it comes standard with presets that allow you to output in a format and size that is easy to pass the LINE Creators stamp examination.
Main features
Japanese prompt: No English required, create illustrations that reflect your intentions with complete Japanese input.
Stamp template: Batch output in default size for LINE, Slack, and Discord.
Refinement function: It is possible to make detailed corrections (color tone, line thickness, facial expression differences, etc.) after generation.
Asset management: Store generated works on the cloud and support team collaboration through shared folders.
Commercial use license: Royalty-free commercial use license is granted, and secondary distribution is also OK.
Technical background
Original learning model: Developed a style-adaptive Diffusion model using over 10,000 commercial stamps from Japanese creators as training data.
High-speed inference engine: A cluster of 8 NVIDIA A100 GPUs achieves an average response time of less than 3 seconds per request.
UX/UI design: Build a drag-and-drop editing screen using React + Tailwind CSS that can be operated intuitively without any design knowledge.
Introduction Example
Major publisher: Created dozens of LINE stamps for book promotion in a short period of time, increasing the number of downloads during the new book campaign by 150% compared to previous versions.
Apparel brand: AI generated banner materials for EC sites, and CTR increased by an average of 1.8 times through A/B testing.
Game company: Used as a rough sketch for character design, reducing creator's brainstorming time by -40%.
Challenges and Future Prospects
Evening out generated variations: To avoid excessively similar output, we are strengthening the prompt understanding diversity algorithm.
Video support: We plan to implement short animation generation from text in the next version.
Collaboration function: We plan to provide a variety of expressions through hybrid collaboration with external AI models (Stable Diffusion, Midjourney). |
| 103 | 豆包（Doubao） | 字节跳动 | China | Multifunctional AI assistant. Supports text, images, audio, video generation, data analysis, online search, etc. | | | | | https://www.doubao.com | DetailsDeveloper company/country
ByteDance/China
Overview
Doubao is a large-scale language model (LLM)-based conversational AI assistant developed by ByteDance, which operates TikTok, with its own research team. A closed beta will begin at the end of 2023, with a public preview available from spring 2024. It has extremely high natural language understanding and generation performance in Chinese, and is used not only for chatbot purposes but also for a wide range of business purposes, such as text creation, product caption generation, and customer service support for e-commerce.
Main features
Conversation engine: Retains context for a long time in multi-turn conversations. Can handle questions and answers, summaries, and small talk naturally.
Automatic text generation: Automatically generate text in a variety of formats, such as EC product descriptions, advertising catchphrases, and SNS posts.
Multilingual translation support: Equipped with a two-way English-Chinese and Chinese-English translation function, and allows for simple proofreading and rephrasing of expressions.
Programming assistance: Provides code sample generation and debugging advice for major languages such as Python and JavaScript.
API platform: We have released an SDK for integrating conversational AI into internal tools and CRM by linking with external systems using REST/gRPC.
Technical background
Base model: Learn a unique LLM with hundreds of billions to trillions of parameters at our own data center. Introducing a large amount of Chinese corpora including web articles, e-commerce information, and video captions.
Optimization: Introducing Knowledge-Augmented Prompting to improve fact-checking accuracy with hybrid browsing that combines a static knowledge base (ByteAtlas) and real-time search.
Infrastructure: With 200 GPUs + AI accelerator hardware, we achieved an average response time of less than 2 seconds per request.
Introduction Example
Major EC platform: Entrusted the generation of product descriptions to Doubao, reducing creation time by 50%. CVR (purchase conversion rate) was +8% compared to previous levels.
Major financial institution: Applied to internal FAQ chatbot to reduce average time to initial response from 3 minutes to 30 seconds. Satisfaction score improved from 4.2 to 4.7.
Educational startup: Built into English learning service to provide essay correction and conversation practice functions. Learning continuation rate improved from 70% to 85%.
Challenges and Future Prospects Suppressing the generation of false information (hallucination): Strengthening external API collaboration for fact checking.
Usage control: For companies, we are developing functions to incorporate industry-specific knowledge and customize interaction policies.
Multimodal: Video and audio generation function (Doubao Vision/Voice) is scheduled to be provided in the second quarter of 2025. |
| 104 | 通义千问（Tongyi Qianwen） | 阿里巴巴集团 | China | A chatbot based on a large-scale language model. It is multilingual and is increasingly being used in business and research fields. | | | | | https://tongyi.aliyun.com | DetailsOverview
Tongyi Qianwen is a large-scale language model service specializing in Chinese developed by Alibaba Cloud. It was first unveiled as part of the "AliGenie" voice AI platform in 2023, and has since expanded its SaaS chat service and API offerings. It has advanced fact-checking and domain adaptation features that are suitable for improving operational efficiency and supporting research and development in large companies.
Main features
Interactive chat: Supports business document summarization, report creation, Q&A, etc.
Domain adaptation: We provide custom models (Fin-Tongyi, Law-Tongyi, etc.) tuned for industries such as finance, legal affairs, and manufacturing.
Multilingual support: Supports international work such as Chinese-English, Chinese-Japanese, Chinese-Korean translation, and cross-lingual summaries.
Knowledge search integration: Link with internal document management system and Alibaba's huge knowledge graph (Aliyun Chuosu) to view the latest information in real time.
AI Workflow: With Workflow Builder (Flow Studio), you can automate internal business processes (task creation → approval → report generation) from chat without any code.
Technical background
LLM scale: approximately 800 billion parameters. Top class in China along with Tencent and Baidu.
Mixed-precision inference: FP8 + INT4 hybrid reduces inference cost by 60% compared to conventional methods.
Privacy protection: Corporate data is captured using proprietary learning within VPC (Federated Learning) to prevent leakage outside the company.
Introduction Example
Major bank: Implemented chat support with customers and increased the number of inquiries handled from 50,000 to 80,000 per month.
Manufacturer: Built an automatic technical document summarization and translation pipeline to improve technology transfer efficiency to overseas subsidiaries by +40%.
Legal office: As a contract review support tool, the initial work time for legal checks has been reduced from an average of 2 hours to 30 minutes.
Challenges and future prospects
Improving the accuracy of industry-specific models: Expanding variations for industries other than finance and legal affairs.
Multimodal enhancement: We plan to announce the "Tongyi Galaxy" concept, which integrates image analysis and voice interaction functions, in 2025. |
| 105 | 文心一言（ERNIE Bot） | 百度公司 | China | A generative AI chatbot specialized in Chinese. The latest model has performance comparable to GPT-4. | | | | | https://yiyan.baidu.com | DetailsOverview
Ernie Bot is a chatbot version of the knowledge-enhancing large-scale language model "ERNIE" developed by Baidu. After the first version of ERNIE was released in 2021, it has been updated successively, and as of 2024, chat functions based on ERNIE 3.0 will be provided via the cloud. Its strengths include collaboration with search engine technology and knowledge graphs, and it achieves highly accurate question answering and text generation while reflecting the latest web information.
Key Features
Knowledge Enrichment Chat: Browse search results in real time and generate responses with the latest news and expert information.
Document summarization/generation: Advanced document creation support, such as long reports, summaries of papers, and proposal draft creation.
Visual Q&A: With the flow of image upload → OCR + image recognition → text explanation, it is possible to generate diagram explanations and explanatory text from product images.
Multimedia commands: Multimodal functions such as audio input/output and video summarization are being gradually expanded.
Enterprise version: A corporate-only plan that supports on-premises deployment and includes access control and data logging functions.
Technical background
Knowledge graph collaboration: Combining Baidu's "Intellectual Brain" knowledge graph and LLM to achieve high-precision question answering through entity relationship analysis.
Dynamic learning: Learn from user feedback in real time and update models every few hours.
Scalability: Supports thousands of simultaneous sessions with in-house POD-based GPU cluster.
Introduction Example
Major media company: Reduced article production man-hours by 30% through automatic generation of article structure ideas and keyword suggestions.
Online education: As a teaching material creation support tool, it automates the teacher's question generation and explanation writing, reducing preparation time per lecture from 2 hours to 45 minutes.
Retail industry: Reduce the initial cost of overseas e-commerce listings by 70% by using automatic translation and summarization of product descriptions.
Challenges and future prospects
Completion of multimodal UX: Enhanced interaction with audio and video is scheduled to be released in the first quarter of 2025.
Custom domain: Promote specialization according to business needs by developing a private model for medical and legal fields. |
| 106 | 混元（Hunyuan） | 腾讯公司 | China | Multimodal large-scale language model. Can generate text, images, and videos. | | | | | https://hunyuan.tencent.com | DetailsOverview
Hunyuan is a multimodal AI platform independently developed by Tencent, and its major feature is that it can understand and generate text, images, audio, and videos in an integrated manner. The first commercial beta version will be released in the spring of 2024, and with repeated updates since then, it has been provided as an API service that can support a variety of internal and external business scenarios. In particular, ``Hunyuan Maestro'' is gaining support for its ability to seamlessly perform a series of workflows, including interpreting complex instructions across multiple modals, transcribing and summarizing text from videos, and even generating short videos from text.
Key Features
Multimodal Context Understanding: Input images, videos, and audio files and automatically summarize and tag relevant text.
Short video generation: Generate videos of several seconds to tens of seconds from text prompts. Speeds up the production of advertisements and SNS content.
Speech synthesis and recognition: Provides natural Chinese speech synthesis (TTS) and noise-tolerant speech recognition (ASR) on the same platform.
Interaction function: Conversation is possible with both text/voice interfaces. You can also define custom skills and turn them into business assistants.
Specialized field tuning: We provide "Hunyuan Blue" and "Hunyuan Health" models that are finely tuned for specific domains such as finance, medicine, and games.
Technical background
Large-scale model: Collaborative learning with multimodal data based on "Hunyuan-1.0" with over 1 trillion parameters.
Hybrid inference: FP8 + INT4 mixed precision calculation reduces inference cost by 60% compared to conventional methods while maintaining low latency.
Distributed training: Utilizes GPU clusters in Tencent's own data center to realize parallel learning of large-scale datasets.
Use Case
E-commerce: Automatically extract features from product images and generate over 100 short advertising videos a day.
Game industry: Automatically generate promotional videos including character movement motions just by prompting.
Online education: Automatically generate screencast videos of lecture slides from text scripts, reducing teaching material production time by 80%.
Challenges and Future Prospects
Stabilizing quality: In particular, improving the accuracy of granularity and aspect ratio control during video generation will be a key development point in the future.
Low latency: A "real-time mode" that reduces API response time to less than 1 second is scheduled to be released in 2025.
Privacy support: Promote on-premises deployment with highly confidential corporate data and minimize the risk of data leakage. |
| 107 | Grok | xAI | America | Conversational AI assistant. Good for creative projects and brainstorming. | | | | | https://x.ai | DetailsOverview
Grok is a conversational AI developed by xAI, an AI research company founded by Elon Musk in early 2024 (formerly known for its "TruthGPT" initiative). It is integrated into the interface of social networking platform X (formerly Twitter), allowing users to engage in intelligent conversations by simply mentioning @Grok in their posts. It is characterized by a style that "maintains a high level of factual consistency while incorporating humor," and it is set apart from other AIs by having guardrails that are strong in filtering out hate speech and misinformation.
Main features
SNS dialogue integration: Can be called directly from tweets or DMs on X.
Fact checking: Automatically detects and warns of misinformation risks by referring to external databases and public sources in real time.
Brainstorming support: Support for generating project ideas, writing outlines, and generating simple code snippets.
Multimedia analysis: When you enter an image URL, you can perform content summarization, tagging, similar image searches, etc.
Reinforcement learning loop: Reflects "useful" and "inappropriate" feedback from users in immediate learning to continuously improve response quality.
Technical background
LLM engine: The independently developed “Grok-1.0” model has hundreds of billions of parameters.
Guardrail design: Introducing a hard constraint layer to suppress hallucination and a policy network to eliminate deflection.
Inference infrastructure: Low latency operation with Amazon AWS Fargate + in-house GPU cluster.
Use Case
Marketer: Used to generate catch phrases and predict performance for SNS campaigns.
Writer: Ask Grok to draft a news article or blog outline.
Customer support: Respond to user inquiries via X in real time and automate the first response.
Challenges and future outlook
Privacy protection: We are currently developing an encrypted storage function for DM conversations.
Platform expansion: We are also considering integration into chat apps other than X and corporate SaaS. |
| 108 | Runway ML Gen-3 | Runway | America | An AI tool for creatives that generates high-quality images and videos from text. | | | | | https://runwayml.com | DetailsOverview
Runway ML Gen‑3 is the latest model provided by the AI creative platform "Runway" that generates high-quality videos from text prompts. It will be released in beta in mid-2024, and is already being adopted by creative industries such as filmmakers and advertising agencies. You can generate a variety of output formats in one stop, from short videos lasting only seconds to seamless VFX effects and stylized looping GIFs.
Main features
Short video generation: Automatically generate HD videos of up to 60 seconds just by inputting text. The scene composition can be specified in the prompt as "Scene 1: .../Scene 2: ...".
VFX Effects Integration: Prompted and real-time compositing of advanced effects such as fire, smoke, and particles.
Video editing assist: Automatically cut and edit uploaded clips, and AI optimizes transitions and audio synchronization.
Style Transfer: Apply various art styles to videos, such as masterpiece style, anime style, minimal design, etc.
Collaboration: Provides project sharing, version control, and commenting capabilities among team members.
Technical background
2D → 3D lattice learning: Equipped with "Latent Video Flow" technology that generates continuous frames with temporal consistency from still images.
UNet extended model: Applying a proven diffusion model for image generation to time series, achieving inference speed within 0.1 seconds per frame.
Cloud GPU: Large-scale distributed processing using a hybrid cluster using Google Cloud TPU v4 and NVIDIA A100.
Use Case
Movie Trailer Creation: Quickly create concept trailers with just a prompt to speed up planning meetings.
SNS advertising: Mass production of dozens of viral loop videos for Instagram and TikTok per day.
Educational videos: Generate illustration-based animations according to the teaching material theme, reducing the time to create video teaching materials by 70%.
Challenges and future outlook
Long video support: The next version of Gen‑3 will support the generation of long sequences of several minutes or more.
Real-time preview: We are currently developing a UI that will allow users to interactively adjust prompts by implementing a preview function during generation. |
| 109 | GitHub Copilot X | GitHub + OpenAI | America | AI coding assistant equipped with GPT-5. Provides real-time code suggestions and debugging assistance. | | | | | https://copilot.github.com | DetailsOverview
GitHub Copilot X is a next-generation AI platform for developers that evolves from the classic Copilot (GPT-4-based code completion). Preview will be released in the second half of 2024, and the chat UI “Copilot Chat” with a built-in GPT‑4o model for ChatGPT, the PR automatic generation function “Copilot for Pull Requests”, and “Copilot Labs” that allows you to write code while interacting in real time on an IDE or terminal will be integrated into one ecosystem. A major feature is that it works with business chats such as Microsoft Teams/Slack, allowing you to complete code reviews and document creation within your workflow.
Main features
Code completion & snippet generation: When you enter a function or class declaration, automatically generates a few lines to hundreds of lines of code instantly.
Copilot Chat (AI Assistant): Displays a chat window in the sidebar of the IDE, where you can ask questions in natural language, such as bug fix advice, algorithm explanations, and refactoring suggestions.
Pull Request automatic generation: Analyze changes and automatically create a PR draft including title, description, and verification steps. One-click integration into your CI/CD pipeline.
Security analysis: Detects vulnerability patterns in code and suggests patch examples. Also implemented license compatibility check for OSS/commercial libraries.
Pair programming support: Automatically generate live coding summaries and key point extractions during remote coding, and provide minutes after the meeting.
Technical background
Multimodal inference: Leverage GPT‑4o to analyze not only text code, but also screenshots and log output.
Edge Cache: Learns the usage patterns of each developer and caches hot code candidates on edge servers for fast response.
Security & Compliance: All corporate code assets are encrypted and stored within a dedicated Azure subscription. Compatible with GDPR and SOC 2.
Use Case
From startups to large enterprises, dramatically improve the productivity of development teams
New employee onboarding: AI guides them in understanding the existing code base, reducing learning costs 60% reduction
Open source project: Activate the community by automatically drafting contribution guides and generating templates for issues
Challenges and future prospects
Suppression of hallucination: Plan to strengthen fact-checking layer based on internal code base to prevent incorrect code generation
Multi-language support: Sequential release of domain-specific models for major languages such as Go, Rust, Kotlin, and Swift
DevOps collaboration: Azure Integrate automatic workflow generation with Pipelines/GitHub Actions to further automate development → testing → deployment |
| 110 | Adobe Firefly | Adobe | America | An AI tool that generates images and videos from text. It is integrated with Adobe products such as Photoshop. | | | | | https://firefly.adobe.com | DetailsOverview
Adobe Firefly is a generative AI platform integrated into Adobe Creative Cloud. It offers a variety of generative tools, including ``Text to Image,'' which generates high-quality images and vector graphics from text input, ``Generative Fill,'' which adds or removes objects from existing images, and ``Text Effects,'' which allows you to freely generate text and layer styles, as well as functions that apply effects to each video frame and automatically generate textures for 3D models. With professional and commercial use in mind, we use licensed data from Adobe Stock for training, ensuring safe and high-quality output.
Main features
Text to Image: Supports multilingual prompts such as Japanese and English. Resolution, aspect ratio, and art style (oil painting, watercolor, manga style, etc.) can be specified in detail.
Generative Fill: Photoshop's "Content-Aware Fill" is powered by AI. Add objects to a specified area, extend the background, and remove unnecessary objects with one click.
Text Effects: When you enter text in Illustrator, professional styles such as 3D text, gradients, and drop shadows are automatically generated.
AI Template: We have prepared a large number of templates such as posters, SNS banners, and web mockups, and you can easily edit materials generated with Firefly by dragging and dropping.
Video compatible: Links with After Effects to automatically generate title animations and transitions of several seconds to tens of seconds.
Technical background
Training data: Utilizes proprietary corpora such as Adobe Stock, Behance, and Typekit. Incorporates Adobe-specific metadata (color code, font information) based on the trained Diffusion model. Cloud inference: Utilizes a distributed GPU cluster on the Creative Cloud runtime to achieve an average generation response time of less than 5 seconds per image.
Security: Generated data is tied to each user's Adobe ID and stored encrypted. On-premises deployment is also possible with the Enterprise plan.
Use Case
Designer: Cut the number of man-hours in half and speed up the cycle from idea sketch to final design
Marketing: Automate the mass production of creatives for posting on SNS and the creation of variations for AB tests
Education: Instantly generate illustrations and visual supplements for teaching materials, reducing lesson preparation time 70% reduction
Challenges and future prospects
Scheduled to release training data transparency report in response to copyright issues
UI/UX optimization of video generation function and expansion of support for long videos
Further strengthening of collaboration with Adobe Express and Lightroom |
| 111 | uPresenter | uLearnbox | international | An educational AI tool that automatically generates presentations and quizzes from text. | | | | | https://upresenter.ai | DetailsOverview
uPresenter is an AI platform for education and training that automatically generates slide materials and quiz questions from text. Combining GPT‑4-based natural language generation with a prompt-driven layout engine, when a user enters the following sentence: ``Create 10 slides about the Third Industrial Revolution and add two review questions to each slide,'' it outputs a set of downloadable materials in PowerPoint format in just a few seconds. Large-scale introduction to educational institutions will begin in 2024, and its use at universities and corporate training sites is rapidly increasing.
Main features
Automatic slide generation: Create visually rich PPTX by simply specifying the theme, number of slides, and font/color style.
Quiz automatic generation: Summarizes the slide content and automatically adds multiple-choice, fill-in-the-blank, and free-answer questions.
Audio narration: Automatically generate scripts for each slide and output video teaching materials with narration using the TTS engine.
Multilingual support: Supports 20 languages including Japanese, English, and Spanish, and creates translated teaching materials all at once.
Analysis report: Collect student answer logs and visualize correct answer rate, answer speed, and comprehension indicators on a dashboard.
Technical background
PPTX engine: Extends the Open XML SDK and uses layout optimization algorithms to automatically adjust the optimal font size and figure placement according to text length.
TTS & video composition: Combined with Amazon Polly/Azure TTS, performs high-quality reading in approximately 3 seconds per slide.
LMS integration: SCORM/xAPI compliant, one-click integration with major LMS such as Moodle and Canvas.
Use Case
University lectures: Teachers take lecture notes with AI , and generate resumes and practice questions all at once
Corporate training: Shorten the launch of e-learning materials from a few hours to a few minutes, allowing for quick updates to the latest information
Personal learning: Automatically generate summaries and confirmation tests for each chapter of specialized books and use them as a self-study support tool
Challenges and future outlook
Design customization: Strengthening the template integration function to match corporate branding
Interactive teaching materials: Interactive video generation that allows you to insert quizzes into the video is scheduled to be released in 2025 |
| 112 | Flux | Black Forest Labs | Germany | A model that generates images from text. Equipped with image editing and compositing functions. | | | | | https://flux.blackforestlabs.ai | DetailsOverview
Flux is the next generation note and knowledge management application for researchers and writers. Markdown-based notes are automatically tagged and summarized, and AI makes suggestions (e.g., diagramming, bibliographic links, action item generation) for users' unfinished notes. Compared to tools such as Obsidian and Roam Research that focus on local storage, Flux implements end-to-end encryption to protect privacy while enhancing cloud sync and collaboration functions. The official version will be released in 2024, and adoption by universities and research institutes is underway.
Main features
Automatic summarization & highlighting: Summarize long notes and highlight key points.
Tagging & link generation: Automatically link related notes to build a knowledge graph in real time.
Charts & Mind Maps: Generate TikZ/Mermaid-based charts and mind maps with a single text instruction.
Co-editing: Provides real-time co-editing and comment/review functions with team members.
Plug-in support: Equipped with plug-ins for two-way collaboration with external tools such as Jupyter Notebook, Zotero, and VSCode.
Technical background
LLM edge integration: Deploy a small LLM (LLaMA 7B, etc.) in the local environment to ensure high-speed processing while ensuring privacy. Offload heavy processing to cloud GPT‑4o backend.
Security: Protect user data with Zero‑knowledge encryption.
Extensibility: Users can define custom scripts and workflows through Flux CLI.
Use case
Academic research: Significantly reduces the number of man-hours required for literature research by summarizing paper reviews and automatically generating critical points
Business documents: Improving work efficiency by structuring meeting minutes and creating tasks at the same time
Creative writing: Supporting ideas for plot structure and character settings
Challenges and future prospects
Local model performance: 13B/30B from 7B model Considering migration to classes and speeding up using quantization technology
Knowledge graph UI: Plans to visualize complex knowledge structures by enhancing 3D display and filtering functions |
| 113 | Windsurf | Windsurf Inc. | America | An AI tool that supports code generation in natural language. Provides an integrated development environment for developers. | | | | | https://windsurf.ai | DetailsOverview
Windsurf is an AI development platform that can automatically generate web applications and dashboards from natural language instructions. The company raised $15 million in a seed round in 2024 and is growing rapidly. Users simply enter the command ``Create a monthly report from sales data and display KPI graphs on the dashboard'' as text, and the backend and frontend will be generated in one go.
Key features
No-code UI design
Automatically place forms, tables, and charts in response to prompts.
Automatic API design
Covering from database schema definition to CRUD endpoint generation.
Real-time deployment
Just push with GitHub integration and deploy immediately to the staging environment.
Security settings
Authority management and OAuth authentication flow are automatically inserted, and MFA/SSO is also supported.
Plugin Ecosystem
We have a wide variety of plugins for major services such as Stripe, Twilio, and Salesforce.
Technical background
LLM integration: Hybrid processing of front-end UI snippet generation and back-end code generation using the GPT-4 model as the core.
Code template: Dynamically recombine the base template of React/Next.js + Node.js/Express to achieve both readability and extensibility.
High-speed inference: Hot-start inference processing with a lightweight container on AWS Lambda, achieving a response within 3 seconds on average.
Use Case
Startup: Launched product MVP in one day and conducted early user testing.
BI team: Rapid prototyping when migrating from Excel to BI tools.
In-house tools: Build small-scale apps for each department, such as sales support and inventory management, one after another without code.
Challenges and future prospects
Custom logic support: Improving automation accuracy of complex business logic and transaction design.
Offline operation: We are currently considering quantization and optimization functions for model execution on edge devices.
Strengthening collaboration: We plan to further automate the workflow for team development (Pull Request → Review → Merge). |
| 114 | SplxAI | SplxAI | Croatia | A tool for evaluating the security of AI systems. Automate vulnerability detection and risk assessment. | | | | | https://splx.ai | DetailsOverview
SplxAI is a cybersecurity assessment platform using AI. Automate system vulnerability scans and penetration tests, and generate security diagnostic reports. It has many implementation results in financial institutions and government agencies, and is unique in that it has a built-in diagnostic framework that is ISO27001 compliant.
Main features
Automatic vulnerability scanning
Web application/API vulnerability detection based on OWASP Top10.
Penetration Testing Module
Simulates advanced attack scenarios such as asset recon, authentication bypass, and privilege escalation.
Risk Assessment Dashboard
CVSS score automatic calculation and security risk prioritization report.
Report Generation & Compliance
Outputs medical certificates in PDF format that comply with various compliance requirements such as PCI-DSS, GDPR, and SOX.
CI/CD integration
Integrated with GitLab/GitHub Actions to run automatic security scans when changing code.
Technical background
AI model: A hybrid operation of an abnormal pattern detection engine using machine learning and a conventional rule-based scanner.
Agentless design: Scans non-intrusively across the network, minimizing impact on operational systems.
Compatible with both cloud and on-premises: Diagnostic data can be kept in either your own AWS environment or on-premises environment.
Use Case
Major securities company: Periodic scanning of trading systems to discover vulnerabilities and shorten the fix cycle.
SaaS provider: Perform automated testing before releasing new features and unify development → security → operation flow.
Government agencies: Used for security evaluation of externally published APIs of critical infrastructure.
Challenges and future prospects
White box diagnosis: Planning to integrate source code level static analysis function.
Zero-day support: Enhancement of machine learning models to improve detection accuracy of unknown vulnerabilities.
Incident response automation: Detection → Notification → Initial response script execution workflow scheduled to be released in the future. |
| 115 | koderAl | koderAl | America | The world's first multi-agent AI coding system. Streamline developer coding work. | | | | | https://www.koderal.com | DetailsOverview
koderAl is a next-generation AI platform that automates coding tasks for developers in a multi-agent configuration. In addition to code completion and debugging support, it has the ability to scaffold the entire project including test code from specifications. It raised $20 million in Series A in 2024, and is currently being implemented primarily for enterprises.
Main Features
Project Scaffold
When requirements are written in natural language, front-end/back-end/infrastructure configurations are output in one package.
Multi-agent cooperation
“Design agent,” “implementation agent,” and “test agent” work together to distribute and process tasks in parallel.
Automatic bug detection and fixing
Analyzes Git history and logs, identifies potential bugs, and suggests fixes.
Knowledge base collaboration
Suggests optimal solutions in conjunction with internal Wiki and Stack Overflow API.
IDE Plugin
An extension for VSCode/IntelliJ that completes code completion and documentation generation within the IDE.
Technical background
LLM: Combination of GPT‑4o + custom fine tuning model.
Agent orchestration: Achieving scale-out with microservices using Kubernetes.
Security: To prevent code asset leakage, prompts and products are fully encrypted & log audited.
Use Case
Financial IT department: Automated refactoring by scanning thousands of lines of existing code in a system renewal project.
Game Development: Quickly build base code for Unity projects and support early prototyping.
Educational platform: Generates automatic scoring test code for practice questions, greatly reducing the burden on instructors.
Challenges and future prospects
Language expansion: We plan to gradually add support for new languages such as Rust, Go, and Swift.
Cost optimization: Aiming to reduce inference cost by using on-premise ML accelerator.
Collaboration: We are currently developing deeper integration functionality with GitHub Copilot and Jira. |
| 116 | HeyGen | HeyGen | America | A video creation tool using AI-generated realistic avatars. Perfect for presentations and customer support. | | | | | https://www.heygen.com | DetailsOverview
HeyGen (formerly known as Movio) is a video generation platform using realistic AI avatars. By inputting a text script, it generates a video narrated by multiple human-like avatars in minutes. It is increasingly being adopted by companies for a variety of purposes, including personalized marketing, e-learning, and internal onboarding videos.
Main features
Avatar library
Preset approximately 50 AI avatars with different age, gender, race, etc.
Speech synthesis
Generate narration with a multilingual TTS engine (English, Japanese, Spanish, and 20 other languages).
Scenario editing
You can adjust the stage, background, BGM, and subtitle display by dragging and dropping.
Personalization
Incorporate user data as merge fields to generate thousands of videos at once.
API integration
You can also embed videos within email campaigns by linking with Salesforce and Marketo.
Technical Background
Deepfake Prevention: Using copyright-free training data and following ethical guidelines.
Cloud rendering: Rendering is distributed and processed on NVIDIA DGX cluster, resulting in an average generation speed of less than 2 minutes per video.
UX/UI: Provides a React-based intuitive dashboard and REST API.
Use Case
Marketing: Mass production of product introduction videos improves CVR by an average of 1.5 times.
Educational institutions: Automatically generate intro videos for online lectures, reducing preparation time by 80%.
HR department: Personalized the onboarding orientation video and improved employee retention rate by +12%.
Challenges and future prospects
Facial expression variations: We are currently improving the motion model to achieve more natural facial expression changes.
Long content: We plan to launch support for video generation of 10 minutes or more in 2025. |
| 117 | Ideogram | Ideogram | America | An AI tool that generates high-quality images from text. Can be used for design and marketing. | | | | | https://www.ideogram.ai | DetailsOverview
Ideogram is an AI tool that generates high-resolution, creative images from text prompts. Public release after closed beta in 2024. In particular, outputs that are both practical and artistic and can be used at work, such as ``concept art,'' ``technical illustrations,'' and ``UI mockups,'' are popular.
Main features
Supports Japanese prompts
Supports multilingual input, allowing detailed instructions in Japanese.
Style Presets
More than 20 preset styles including cyberpunk, mid-century, and minimal design.
Parameter adjustment
Fine tuning of composition, color, light and shadow, resolution, etc. using the GUI.
Batch generation
Generate up to 50 sheets at once from the same prompt and use them for A/B testing.
Version Management
Manage generation history on timeline and rollback to past versions.
Technical background
Diffusion model improvement: Based on Stable Diffusion, implements a unique acceleration layer that achieves both memory efficiency and generation speed.
Cloud GPU: Utilizes Google Cloud TPU v4 to achieve output within 4 seconds on average.
Safety filter: In order to eliminate copyright risks during commercial use, learning data is limited to copyright-cleared materials such as Adobe Stock.
Use Case
Game Designer: Use it as a basis for creating concept art and shorten the creative cycle by 50%.
UI/UX designer: Generate several visual plans for prototyping at once to streamline customer proposals.
Marketing: Generate a large amount of A/B test materials for SNS advertising visuals.
Challenges and future prospects
Multimodal expansion: Video support and 3D model output will be supported in the future.
Team collaboration: We are planning to implement real-time collaborative editing function in the first half of 2025. |
| 118 | Afforai | Afforai | America | An AI research tool that extracts insights from vast amounts of data and documents. Useful for research and market analysis. | | | | | https://www.afforai.com | DetailsOverview
Afforai is an AI research platform that automatically extracts insights from large datasets and documents and generates research reports. By reading papers, technical documents, web content, in-house materials, etc. all at once, you can output summaries of key points, trend analysis, graph creation for hypothesis verification, and Q&A-style reports in a few minutes. It is increasingly being introduced in finance, pharmaceuticals, and tech startups as a service for professionals such as data scientists, marketing researchers, and strategy consultants who want to quickly get a bird's-eye view of vast amounts of information.
Main features
Document ingest: A variety of sources such as PDF, Word, HTML, and CSV formats can be uploaded in bulk.
Automatic summary & key finding: Extract only key messages and conclusions from dozens of pages of reports and organize them in bullet points.
Trend analysis dashboard: Automatically generate word clouds, time series graphs, and correlation matrices based on text mining.
Q&A interface: When asked in natural language to the imported materials, ``What are the main issues of this project?'' and ``Please tell me the main points of competitive comparison,'' the AI extracts and answers the relevant parts.
Custom template: PDF/PowerPoint output that reflects a company's unique report format and branding is possible.
Technical background
Using LLM: Using GPT-4o and our own fine-tuning model, we achieve highly accurate summarization while preserving the context of long texts.
Knowledge graph integration: Automatically visualize relationships between materials and topic links using an internal knowledge graph.
Distributed processing: Microservices architecture on AWS supports scale-out analysis of dozens of large-scale documents at the same time.
Introduction Example
Major pharmaceutical company: Aggregated clinical trial reports and academic papers into Afforai, shortened key review of new drug development projects from 2 weeks to 2 days.
Financial analysts: Automatically analyze market reports and financial statements, reducing the time required to draft investment decision reports from scratch by 80%.
Marketing startup: Collectively collect and summarize competitive information on the web and automatically generate weekly business strategy meeting materials.
Challenges and future prospects
Compatibility with technical terms: There is an urgent need to improve the accuracy of interpretation of industry-specific technical terms and abbreviations. We are currently implementing custom dictionary functions for each industry.
Enhanced interactivity: Real-time collaborative editing mode and filtering and bookmarking functions within the dashboard are scheduled to be released in Q2 2025.
Privacy protection: We are creating an environment where you can choose between on-premises deployment or processing within VPC, and can operate with confidence even when analyzing highly confidential materials. |
| 119 | Voicenotes | Voicenotes | America | An AI tool that instantly converts audio input into structured notes. Convenient for recording ideas and meeting notes. | | | | | https://www.voicenotes.ai | DetailsOverview
Voicenotes is an AI tool that records audio conferences and face-to-face interviews in real time and instantly converts them into structured notes (minutes). When recording data is uploaded to the server, conversation speaker separation, statement summarization, and action item extraction are automatically executed. Furthermore, it is possible to generate a PDF report that summarizes key points in bullet points and automatically send it to Slack or email. It streamlines all "speaking → recording" workflows such as sales meetings, interviews, counseling records, etc.
Main features
Speaker separation & timestamp: Identify the voices of multiple speakers and record each utterance with a timestamp.
Automatic summary: Generates a summary of important points from long conversations into 3-5 lines.
Action item extraction: Create a list of tasks and next action items, and automatically tag them if the person in charge and due date information is included.
Multilingual support: Supports voice recognition in 10 languages including English, Japanese, and Spanish.
Integrated workflow: Direct integration with Zoom, Teams, and Google Meet, sharing to Dropbox and Google Drive, automatic email sending function.
Technical Background
ASR Engine: A hybrid approach based on our proprietary noise-tolerant automatic speech recognition model and Kaldi.
NLP pipeline: After converting speech to text, parallel processing of summarization and task extraction using a Transformer-based model.
Real-time processing: Utilizes WebSockets to provide real-time summary previews during meetings.
Introduction Example
Major human resources company: Entrusted recording of telephone interviews to report creation to Voicenotes, reducing input time for recruitment staff by 100%.
Market research company: Automatically processes a large amount of interview audio sources and completes analysis report creation in a few days instead of a few weeks.
Healthcare startup: Records and summarizes counseling audio to streamline recording work for doctors and coaches.
Challenges and future outlook
Improving accuracy: We are currently developing a custom dictionary function to improve the recognition accuracy of technical terms and proper nouns.
Interactive editing: We plan to introduce a collaboration function in Q2 2025 that allows you to modify and approve the summary content on the chat UI. |
| 120 | EmoBay | EmoBay | Hongkong | AI platform for mental health support. We provide a chatbot that is available 24 hours a day. | | | | | https://emobay.org | DetailsOverview
EmoBay is an AI chatbot platform specializing in mental health support. Using custom LLM designed based on psychological and clinical data, we provide "self-care AI" that allows you to consult 24 hours a day, 365 days a year. Users can input their feelings and concerns via text or voice and receive worksheet presentations and stress management advice based on empathic responses and cognitive behavioral therapy (CBT). It is increasingly being introduced as an EAP (Employee Assistance Program) by companies and as a school counseling aid by educational institutions.
Key Features
Emotion Detection & Empathetic Response: Estimates emotions from text input and generates empathetic messages with the appropriate tone.
CBT Worksheet: Automatically creates task setting and cognitive restructuring worksheets in response to user requests.
Regular mental check: Answer daily/weekly self-check questions and visualize your score on the dashboard.
Anonymous group support: Generate anonymous chat rooms between users with the same issues to facilitate peer support.
Emergency navigation: When suicidal thoughts or crisis signals are detected, hotline numbers and medical institution information are immediately displayed.
Technical background
Custom model: Uses specialized LLM with medical/psychological literature and clinical session records as training data.
Privacy compliance: HIPAA/PIPEDA compliant data encryption and anonymization processing.
UX design: Adopts emotional design and provides a sense of security with a UI with character.
Installation Example
Major IT company: Integrated EmoBay into its EAP program and increased employee usage rate by 70% compared to previous levels.
University campus: Introduced as a mental health check tool for students and serves as an initial consultation shelter before applying for counseling.
Challenges and future prospects
Multilingual support: Quality improvements and localization of Japanese, Korean, and Spanish versions are scheduled to be released sequentially in 2025.
Expert collaboration: Under the supervision of clinical psychologists and psychiatrists, we plan to obtain third-party certification to objectively evaluate AI response accuracy. |
| 121 | Vastav AI | Zero Defend Security | India | An AI tool specialized in deepfake detection. Determine the authenticity of media with high accuracy. | | | | | https://www.zerodefendsecurity.com | DetailsOverview
Vastav AI is an AI platform specialized in deepfake detection. We provide a service that highly accurately determines whether videos and images have been tampered with and guarantees the authenticity of the content. It combines pixel-level tamper detection models powered by machine learning with metadata analysis and blockchain-based trail management, and is increasingly being adopted by media companies, law enforcement agencies, and social platforms.
Main features
Deepfake detection engine: Captures minute changes in facial parts and audio features in videos, achieving a fraud detection rate of over 98%.
Metadata analysis: Analyzes file headers and encoding information to detect signs of tampering.
Blockchain audit trail: Detection results are issued in NFT format and can be used for evidence preservation.
API & SDK: The platform is provided as a REST API and can be easily integrated into mobile apps and web services.
Automatic report generation: Outputs a detailed report summarizing detection results in PDF/HTML format, and can also be used to submit legal evidence.
Technical Background
Multi-model ensemble: Integrates CNN-based pixel anomaly detection model, audio spectrogram analysis model, and on-chain metadata matching model.
Real-time processing: We also offer a lightweight model optimized for edge devices and also supports streaming video.
Continuous learning: Automatically collect newly emerging deepfake methods and continuously update the model through online learning.
Case study
Global SNS company: Detects deepfakes in uploaded videos in real time to prevent the spread of misinformation.
News media: Assess the authenticity of news footage and contribute to preventing fake news.
Government agencies: Used to detect tampering with election-related videos for early detection of fraudulent operations.
Challenges and future prospects
Compatible with new methods: Accelerating adaptation to GAN3.0 and Diffusion-based tampering technology, always maintaining detection accuracy above 98%.
Regulatory compliance: We have established operational guidelines that comply with each country's digital evidence laws and privacy regulations (GDPR, CCPA). |
| 122 | Claude 3.5 Sonnet | Anthropic | America | An AI chatbot that excels in analyzing long texts and making nuanced inferences. Emphasis on safety and consistency with human values. | | | | | https://www.anthropic.com | DetailsOverview
Claude 3.5 Sonnet is a model specializing in dialogue and reasoning that is part of the latest generation large-scale language model "Claude" series developed by Anthropic. Released in early 2025, it offers significant improvements over the previous version (Claude 3) in three areas: consistency of long documents, advanced reasoning tasks, and compliance with safety guidelines. It is especially used as an assistant for businesses, from responding to FAQs to generating risk analysis reports.
Main functions
Advanced reasoning
Read contracts and specifications containing thousands of tokens at once and automatically extract key points and highlight risk areas.
Multi-turn dialogue
Holds up to 10,000 tokens of conversation context for long-term project consultation and brainstorming.
Safeguard
Equipped with output filtering (exclusion of biased words and harmful content) according to the user's expertise and compliance requirements.
Custom knowledge integration
Link with internal intranet documents and DB to generate responses by referencing the latest data in real time.
Multimodal input
Input not only text but also PDF, images, and spreadsheets, allowing answers in mixed contexts (beta version).
Technical background
Model architecture: Combining a Transformer-based model with Anthropic's proprietary "Synthesized Chain-of-Thought" technology to internally visualize and verify the inference process.
Safety Framework: In addition to RLHF (Human Feedback with Reinforcement Learning), we introduced a policy-driven self-auditing loop called "Constitutional AI" to block over 98% of harmful outputs.
Infrastructure: Using in-house GPU clusters in conjunction with AWS Inferentia, we achieved an average response time of less than 500ms.
Use Case
Legal department: Automate contract draft summarization and amendment proposals, reducing review man-hours by 60% compared to conventional methods.
Research and development: One-stop summary creation of technical paper summaries and related research, reducing literature research time by 75%.
Customer support: Automated FAQ response based on past inquiry history, shortened initial response time from 5 minutes to 1 minute.
Challenges and Future Prospects
Long-term memory function: We are planning to strengthen persistent session management across documents containing tens of thousands of tokens.
Multimodal enhancement: Integrate speech recognition/synthesis and plan to deploy it to voice assistants in the second half of 2025. |
| 123 | Google Gemini 1.5 | Google | America | Multimodal AI that integrally processes text, images, and videos. Used for research and automation. | | | | | https://gemini.google.com | DetailsOverview
Google Gemini 1.5 is a multimodal large-scale language model developed by Google's AI research division "Brain." Compatible with a variety of modals such as text, images, audio, and video, 1.5, released in early 2025, surpasses the previous version in three areas: "context understanding," "code generation accuracy," and "creative generation performance." Notebook integration for researchers and Vertex AI collaboration for enterprises have been enhanced.
Main features
Multimodal response
Upload an image → Answer questions such as "Tell me the main points of this chart" using numerical analysis of the chart and natural sentences.
Code Assist
Provides Python, SQL, and TensorFlow code generation and debugging assistance on Notebook.
Creative Studio
One-stop text → image/video generation, design material creation, and animation editing using voice commands.
Large-scale search integration
Reference Google search results in real time and generate FAQ answers and reports based on the latest information.
Security for enterprises
We provide an enterprise plan that safely handles corporate data by supporting VPC Service Controls and linking IAM permission settings.
Technical Background
Architecture: Dynamic model calling using the Mixture-of-Experts (MoE) layer covers multi-domain tasks while keeping the number of parameters down.
Continual Learning: Regularly incrementally learns the latest information from Google searches and YouTube to maintain data freshness.
Inference platform: Optimize scale-out inference on Google Cloud AI Platform using TPU v5.
Use Case
Advertising agency: Completely automated from generation of image ideas for advertising creatives to text generation for A/B testing.
Research institutions: Graph analysis of experimental results → automatic report creation → paper draft generation completed within Notebook.
Manufacturing industry: Automate loading of blueprint images → parts list extraction → purchase order draft generation.
Challenges and future prospects
Explainability (XAI): We plan to enhance visualization tools that can trace the basis of answers.
On-device inference: We plan to provide a quantization model for Edge TPU to achieve high-speed response in mobile/IoT. |
| 124 | Midjourney v6 | Midjourney | America | An AI tool that generates high-resolution, artistic images. Popular with designers and artists. | | | | | https://www.midjourney.com | DetailsOverview
Midjourney v6 is the latest version of Midjourney, a text-to-image generation platform for creatives. Released in 2025, resolution, detail expression, and artistic color and texture reproduction will be dramatically improved. Leveraging the community's Discord interface, prompt history and variation generation workflows are intuitive.
Main features
High-definition rendering
Supports a variety of styles from real photos to oil paintings at 8K resolution.
Texture & Lighting
Supports detailed texture expressions (fabric texture, metal reflections) and complex lighting simulations.
Variation/Upscaling
Select your favorite candidate → Partial regeneration or super-resolution upscaling is possible.
Community Gallery
Use samples of other users' prompted works with the share and remix feature on Discord.
License Commercialization
The product has commercial use rights (included in the subscription plan) and is copyright cleared.
Technical background
Diffusion improvements: Equipped with next-generation diffusion model "Mid-Diffuse v2" and optimized noise removal filter and color grading.
Infrastructure: Combining AWS EC2 GPU cluster and in-house cache layer, prompt to image generation takes less than 6 seconds on average.
UX design: By using Discord Bot as the front end, anyone can use it with low learning cost.
Use Case
Game development: Speed up the rough creation of concept art and background illustrations for character designs.
Advertisement production: Generate a large number of campaign visual proposals and speed up internal reviews.
Art project: Used to generate NFT art and create visual concepts for installations.
Challenges and future prospects
Prompt EQ: A "prompt analysis tool" that visualizes the correlation between prompts and results is scheduled to be released in Q3 2025.
Video generation: Still image → video conversion function currently under development for release in 2026. |
| 125 | Suno AI | Suno AI | America | An AI tool that generates copyright-free background music. For content creators. | | | | | https://www.suno.ai | DetailsOverview
Suno AI is a music generation platform for creatives that automatically generates copyright-free background music, sound effects, and soundtracks from text prompts or melody input. After its official release in the second half of 2024, adoption has rapidly increased in podcasts, YouTube videos, game development sites, etc. The feature is that you can specify the style (jazz, rock, synthwave, etc.), instrumentation, tempo, and mood in detail.
Main features
Text prompt generation
Output a few minutes of music just by specifying things like "calm piano solo" or "intense synth bass + drums".
Melody Completion
Analyzes MIDI melodies uploaded by users and automatically generates accompaniment and arrangements.
Loop & Stem Output
Supports export of loops/stem (track by track) that are easy to use as BGM material.
Mastering
AI-based mastering process automatically performs volume equalization and equalization.
Commercial License
Grants a commercial use license for individuals and corporations without additional fees.
Technical background
Music generation model: Hybrid architecture that combines a DiffWave-based expansion model with Transformer.
Large-scale music corpus: Tens of thousands of public domain and licensed songs are used as training data.
Low latency inference: Realizes music generation in around 10 seconds using GPU cloud.
Use Case
Content creator: Quickly source theme music and BGM for YouTube/Podcast.
Game development: Instantly generate sound effects and BGM for prototypes to shorten the development cycle.
Advertisement production: Generate soundtrack ideas for commercials and promotional videos all at once.
Challenges and future outlook
Copyright risk reduction: Publish learning data transparency report and respond to legal audits.
Live performance mode: Real-time JAM session generation function scheduled to be launched in 2025. |
| 126 | Notebook LM | Google | America | An AI tool that summarizes long blogs, PDFs, and YouTube videos and turns them into audio podcasts. | | | | | https://notebooklm.google.com | DetailsOverview
Notebook LM is an "AI notebook assistant for researchers" announced by Google in 2024. When you import a variety of document sources, including text, PDFs, YouTube videos, and web pages, into Notebook, AI helps you summarize, generate code snippets, interpret experimental results, propose your next research question, and more. It follows a Colab-style interface and can be easily integrated into existing research workflows.
Main features
Document summary
Cross-reference multiple sources to comprehensively summarize research background and gaps.
Code Generation
Generate Python/R/SQL experiment scripts and run them directly within Notebook.
Visualization
AI automatically analyzes data frames → generates matplotlib-based graphs and displays them inline.
Research Q&A
If you ask questions such as "How should we verify the validity of this result?" and "What are related studies?", we will respond with links to references.
Collaboration
Equipped with Google Drive synchronization, sharing permission settings, and comment/review functions to support collaborative writing by research teams.
Technical background
Model integration: Adopts PaLM 2 + Gemini API as backend to achieve both large-scale knowledge base and high-speed inference.
Security: Confidential research data is processed within VPC Service Controls, and access logs are managed with Auditor.
Data-driven learning: Anonymize user operation data on Notebook and feed it back to learning to continuously improve functionality.
Use Case
University research lab: Complete the entire process from paper survey → experimental design → result verification → paper draft in Notebook.
Corporate R&D: Automate patent document analysis and demonstration experiment script generation to shorten new project launch time by 50%.
Challenges and future prospects
Large-scale data processing: Improving memory management and speedup when handling multi-GB data sets.
Multimodal support: Experimental analysis support function for image and audio data will be added in 2025. |
| 127 | Omi | Omi | America | A wearable AI device that records and transcribes conversations and generates summaries and action items. | | | | | https://www.omi.ai | DetailsOverview
Omi is a revolutionary AI solution that completes the "recording → transcription → summary → action item generation" of meetings, business negotiations, and interviews using a wearable device. Users can select their preferred style, such as a necklace or clip type, and simply attach the device to automatically capture all interactions. More than 20 languages are supported, and text conversion and summaries are reflected in your smartphone app or PC dashboard in real time through cloud synchronization.
Main features
Real-time transcription: Using high-performance microphones and noise canceling technology, each speaker's speech is automatically converted into text.
Automatic summary generation: Convert hours of audio data into a 5-7 line bulleted summary in 5 seconds.
Action item extraction: Extracts and creates a list of "what to do next" from the content of the statement with pseudo tags for the person in charge.
Sentiment analysis: Analyzes the tone and emphasis points and graphs the atmosphere (positive/negative) of the meeting.
Cloud & On-Premise Deployment: Custom deployment to private cloud environments is also supported to protect confidential information.
Technical Background
ASR Engine: A hybrid of an in-house developed neural network-based speech recognition model and the Kaldi framework.
NLP pipeline: Fine-tuned the Transformer-based summary model and adopted a hybrid of rule-based and machine learning-based for action extraction.
Wearable optimization: Links ARM Cortex-M processor + DSP accelerator to perform noise reduction processing on the device.
Use Case
Sales Department: Record customer visits and business negotiations with Omi, reducing the number of missed interviews and follow-ups to zero.
Consulting: Automatically convert client workshops into reports and reduce delivery document creation time by 70%.
Research survey: The audio data of the interview is analyzed in bulk, and a draft of the research report is automatically generated by AI.
Challenges and future outlook
Support for long-term recording: Enhanced battery durability and data offload function to enable continuous recording for 8 hours.
Multilingual summarization: Fine tuning for each language is currently underway to improve the accuracy of summarization in languages other than English and Japanese.
Cooperative development: We aim to fully automate the process of meeting → task creation by strengthening API collaboration with CRM and project management tools. |
| 128 | Leonardo AI | Leonardo AI | America | A design support tool that uses AI to generate logos, marketing materials, and custom illustrations. | | | | | https://www.leonardo.ai | DetailsOverview
Leonardo AI is an AI image generation and editing platform specialized for creative and product design. Upload a stylus or rough sketch to instantly generate multiple design variations. We provide a one-stop service for creating 3D model textures, UX/UI mockups, and automatic animated GIF creation, dramatically speeding up prototyping for design teams.
Main features
Rough sketch input: Scan handwritten illustrations → AI formats and complements them, converting them into commercial quality illustrations.
Style Export: Apply specified art styles (watercolor, pencil, photorealistic, etc.) with one click.
3D texture generation: Automatically generate PBR textures (albedo, roughness, normal maps, etc.) for Blender and Unreal Engine.
UI/UX mockup: Export JSON for Figma when you specify a text such as "Create a simple authentication screen."
Livestock Collaboration: Host real-time collaborative editing sessions and brainstorm generated images across teams.
Technical background
Diffusion + GAN hybrid: A unique architecture that combines the stability of Diffusion and the sharpness of GAN inside the model.
Cloud rendering: Using AWS EC2 GPU + in-house host GPU cluster, we achieved high-speed generation with an average of 4 seconds per request.
Metadata management: Generation history management function with generation parameters ensures version control and reproducibility.
Use Case
Game development: Responsible for creating concept art for prototyping and initial drafts of character textures.
Product design: Before 3D rendering of a new product, we quickly create appearance plans and color variations.
Advertising agency: Shorten the feedback loop from rough generation of campaign visual plans to correction instructions.
Challenges and future outlook
Video motion generation: Automatic rigging from 2D to 3D and short animation generation function scheduled for beta release at the end of 2025.
On-premises version: Private deployment is currently being prepared for large companies that cannot leave confidential data outside the company. |
| 129 | Bolt.new | Bolt.new | America | An AI platform that allows you to prototype, edit, and deploy applications without coding. | | | | | https://bolt.new | DetailsOverview
Bolt.new is a no-code development platform that utilizes AI, and is characterized by "generating dynamic web applications with a single prompt line." Internally, it is built seamlessly using GPT-4o, the front end is React + Tailwind, the back end is Node.js + Express, and the infrastructure is Serverless Framework. Complex processes such as database settings, authentication functions, and external API collaboration are fully automated, and idea validation (PoC) can be completed within an hour.
Main features
Prompt-driven development: When you instruct in natural language, such as "Create user registration function and profile screen," code is automatically generated.
Authentication & Authority Management: Automatically implements JWT-based authentication and allows access control to be set by user group or role.
Database migration: Automatically detect schema changes, generate and apply migration scripts.
CI/CD integration: Set up Push → Automated testing → Deployment to AWS Lambda with no code using GitHub integration.
Plugins & Add-ons: Standard template plugins for major services such as Stripe, Twilio, and SendGrid.
Technical background
Prompt template creation: Prompt presets are prepared for each industry and application to ensure generation accuracy and reproducibility.
Architecture coding: Equipped with a post-processing engine that analyzes the code generated by GPT by AST and refactors it into a safe and optimal structure.
Serverless optimization: Make each API endpoint an independent function to enable scalable operation with minimal cold start.
Use Case
Startup: Launch MVP development with Bolt.new and complete investor demo in a few hours.
In-house tools: Quickly deploy small-scale apps tailored to departmental needs, such as sales support apps and material inventory management tools.
Hackathon: Featured as a platform where anyone can concentrate on implementing ideas regardless of their technical ability.
Challenges and future prospects
Support for complex logic: We plan to enhance the automatic generation of advanced functions such as transaction management and batch processing at low layers.
Automatic document generation: A function that uses AI to output API references, ER diagrams, and flowcharts is scheduled to be released in the first half of 2025 |
| 130 | Ernie Bot | Baidu | China | A large-scale language model developed by Baidu. Version 4.5 released in March 2025. | | | | | https://yiyan.baidu.com/ | DetailsOverview
Ernie Bot is a chatbot version of the knowledge-enhancing LLM "ERNIE" series developed by Baidu. It will be commercially released in 2024 based on "Ernie 3.0 Titan" and integrates web search, knowledge graph, and speech synthesis/recognition. In addition to native Chinese speakers, it also supports multilingual dialogue such as English and Japanese, and is rapidly expanding mainly in the B2B market.
Main features
Real-time search integration: Integrates with Baidu search API to generate answers based on the latest news and statistical data.
Knowledge Graph Q&A: Explicitly explain relationships between entities using Baidu's "Intellectual Brain" knowledge graph.
Multilingual translation/summarization: Provides one-stop bidirectional translation between Chinese-English, Chinese-Japanese, and Chinese-Korean, as well as long document summarization functions.
Multi-modal support: Image upload → OCR + image content description, voice input → text conversion → answer, video summary (β version).
Enterprise version: A corporate plan with on-premises deployment, access control, and log auditing features.
Technical background
Parameter scale: ERNIE 3.0 Titan with approximately 1,000 billion parameters is the core, achieving both high-speed inference and low latency.
Self-learning loop: Incorporates user feedback and continuously updates the model over hours.
Large-scale cloud infrastructure: Leverages Baidu Cloud's GPU cluster and Ascend AI processor to support 3,000 simultaneous sessions.
Use Case
Financial institutions: Used for investment report summary chatbots and automatic summarization of corporate financial results briefing materials.
Media: Reduce writing time by 30% by automatically creating article structure drafts and extracting key points.
Education: Enhance learning support with automatic subtitle generation and summaries of online lectures, and Q&A functions.
Challenges and future prospects
Ethical guidelines: Strengthen compliance with the "AI Charter" to suppress bias and reduce the risk of generating false rumors.
International expansion: Official release in Japanese and English speaking countries and enhanced localization planned for 2025. |
| 131 | MiniMax | MiniMax | China | We have developed the multimodal LLM "MiniMax-01" and the voice model "Speech-02." | | private | | API、Web | https://www.minimax.io/ | DetailsOverview
MiniMax is a multimodal AI platform originating from China, with a large-scale language model called "MiniMax‑01" and a voice interaction model called "Speech‑02" at its core. A beta version for businesses will be released in 2024, and it can be widely applied not only to text summarization and dialogue, but also to general business automation, such as AI call centers, voice assistants, document OCR + dialogue flow generation, etc.
Key Features
Text Interaction & Summarization: Thousands of tokens of document summarization, Q&A with high accuracy.
Voice dialogue service: Automates everything from IVR (interactive voice response) to manned transfer by linking with telephone API.
Document OCR + Chat: Search and answer questions using OCR of PDF and imaged documents → convert to text → chat format.
Multimodal generation: Text → image, text → audio, and image → caption generation are performed in an integrated manner.
Enterprise integration: Seamless integration with chat tools such as WeChat, Enterprise WeChat, Slack, and Teams.
Technical background
Base model: MiniMax‑01 with 120 billion parameters trained in our own data center.
Speech model: Speech‑02 is an improved version of the Transformer‑TTS+Tacotron system, and its sound quality and speech naturalness are among the top in the industry.
Privacy: Internal data is exclusively learned in VPC, and output content is automatically saved as an audit log.
Use Case
BPO company: Automated AI response for 50% of call center operations, reducing costs by 30%.
Manufacturing industry: Build an OCR + summary → automatic report generation workflow for on-site reports, reducing reporting time by 90%.
Medical institutions: Link medical literature summaries and interview chatbots to support doctors' preparations before their first consultation.
Challenges and Future Prospects
Multilingual enhancement: Improvements in Japanese and English dialogue/summary accuracy, and technical terminology support scheduled to be released in Q1 2025.
Real-time collaboration: We are currently developing a "meeting mode" that summarizes and extracts tasks from multiple users' simultaneous conversations. |
| 132 | Auto-GPT | Toran Bruce Richards | Unknown (OSS) | autonomous agent | | Free (OSS) | | open source, CLI | https://github.com/Significant-Gravitas/Auto-GPT | DetailsAn agent framework that automatically executes GPT-4 continuously. Specialized in task automation. |
| 133 | ChatSonic | Writesonic | India | Conversational AI, text generation | | Free/Paid | | Web apps, APIs | https://writesonic.com/chat | DetailsWritesonic's ChatGPT alternative chatbot. News updates and image search integration are possible. |
| 134 | Luma AI | Luma AI | America | 3D generation, image processing | | Free (with limitations) / Paid | | Web, mobile app | https://www.luma.ai | DetailsA research-oriented tool that generates realistic 3D models from text. |
| 135 | AWS SageMaker | Amazon Web Services | America | machine learning platform | | Pay-as-you-go system | | Cloud service (SaaS) | https://aws.amazon.com/sagemaker | DetailsMachine learning platform provided by Amazon. Model building, training, and deployment all in one go. |
| 136 | Azure Machine Learning | Microsoft | America | machine learning platform | | Pay-as-you-go system | | Cloud service (SaaS) | https://azure.microsoft.com/aml | DetailsMicrosoft's cloud ML service. Equipped with a wealth of AutoML and MLOps functions. |
| 137 | AutoKeras | Texas A&M University | America | AutoML | | Free (open source) | | Python library (OSS) | https://autokeras.com | DetailsOpen source AutoML library. Automatic exploration of deep learning models is possible. |
| 138 | Hugging Face Hub | Hugging Face | America | Model sharing, LLM | | Free / paid plans available | | Web, API, Python libraries | https://huggingface.co | DetailsOver 25,000 pre-trained models published. Ecosystem including Transformers/Diffusers. |
| 139 | Salesforce Einstein | Salesforce, Inc. | united states of america | AI built into Salesforce CRM. Automate predictive analysis and recommendations. | | | | | https://www.salesforce.com/einstein | DetailsOverview
Salesforce Einstein is an AI layer built into the global CRM platform Salesforce, seamlessly delivering AI capabilities to all operations, including sales, marketing, and customer service. The core of Einstein is predictive analytics using machine learning, and natural language processing (NLP) and image recognition functions using deep learning. We provide a one-stop solution for a variety of use cases, including ``predicting the probability of closing a deal,'' ``optimal email sending timing,'' ``first response using a chatbot,'' ``sentiment analysis of social media posts,'' ``photo-based asset management,'' and ``summarizing business negotiations and extracting next actions.'' Its greatest strength is that all functions can be configured without any code (clicks and points), and it has a wide range of implementation results from small to medium-sized businesses to Fortune 500 companies.
Main features
Einstein Lead Scoring: Automatically analyzes lead attributes and behavior data and automatically ranks leads that are likely to convert into business negotiations.
Einstein Opportunity Insights: Real-time analysis of opportunity history and activity logs to alert you to early closing signs and risks.
Einstein Bots: Build chatbots based on CRM data with no code. Achieved automation of FAQ response and case filing.
Einstein Vision & Language: Supports product shelf checking using image recognition, automatic report summarization and sentiment analysis in natural language.
Einstein Prediction Builder: AutoML feature that allows even non-engineers to create their own predictive models with drag and drop.
Benefits of implementation
Improved productivity: Automate sales staff's prospect evaluation work, reducing man-hours by hundreds of hours per year.
Improved accuracy: Use machine learning to extract customer behavior patterns that are often overlooked manually, increasing conversion rates by an average of 15%.
Cost optimization: Hybrid architecture with OpenAI and in-house ML environment provides high-speed inference while controlling cloud costs.
Technology infrastructure
Infrastructure: Globally distributed deployment on Salesforce Hyperforce.
Security: HIPAA/GDPR/SOC2 compliant, data is always encrypted and stored within the same region.
Extensibility: Can also be linked with external data via Heroku or MuleSoft. |
| 140 | SAP AI Core | SAP SE | Germany | AI platform for SAP business processes. Perform advanced analysis in conjunction with ERP data. | | | | | https://www.sap.com/products/ai.html | DetailsOverview
SAP AI Core is an enterprise AI foundation that is deeply integrated with SAP business applications such as ERP and CRM. Capture operational data from manufacturing, retail, logistics, finance, human resources, and more in real time and apply it to a variety of scenarios, including demand forecasting, inventory optimization, fraud detection, preventive maintenance, and employee turnover risk analysis. We offer a full stack from AutoML to MLOps, centralizing model training, deployment, monitoring, and lifecycle management. Linkage with business processes is completed on SAP Business Technology Platform (BTP), enabling AI operations that meet enterprise governance and compliance requirements.
Key Features
AutoML Pipeline: Automate model creation with no code. Covers data preprocessing, feature engineering, and hyperparameter optimization.
MLOps: Comes standard with GitOps-based continuous delivery capabilities and supports safe model releases and rollbacks to production environments.
Real-time inference: Stream event data from SAP S/4HANA and SAP Commerce Cloud to achieve low-latency inference.
Explainable AI: Visualizes the basis of prediction and feature importance for each model to support decision-making by business personnel.
Enterprise governance: Equipped with data access control, audit log, and privacy mask functions, and can operate in compliance with GDPR/ISO27001.
Advantages of implementation
Process optimization: Predictive maintenance features reduce unplanned downtime by 25%.
Inventory management: Improved inventory turnover rate by 20% with a demand forecasting model that utilizes sales and procurement data.
Human resources: Reduced turnover rate by 10% through human resource allocation optimization model.
Technology platform
BTP integration: Links with SAP Data Intelligence and SAP Analytics Cloud to seamlessly realize data lake → analysis → visualization workflow.
Cloud compatible: Also available on AWS, Azure, and Google Cloud. |
| 141 | Oracle Fusion AI | Oracle Corporation | united states of america | AI capabilities integrated into Oracle Cloud Applications. Supports predictions, chatbots, document summaries, etc. | | | | | https://www.oracle.com/cloud/fusion | DetailsOverview
Oracle Fusion AI is a suite of AI capabilities built into Oracle Cloud Applications (ERP, HCM, CX, SCM, etc.). We provide modular services such as machine learning-based "predictive analysis", "chatbot" using natural language dialogue, "Content Intelligence" for document summarization and classification, and "Visual Builder" for image and video analysis. These operate on Oracle's integrated data lake (OCI Data Lake), enabling seamless data distribution and secure operation. Utilizing Oracle's strength in large-volume data processing infrastructure, we will optimize corporate operations on a global scale with AI.
Main features
Predictive Planning: Perform advanced predictive analytics such as demand forecasting, price optimization, and cash flow forecasting with no code.
Digital Assistant: A customizable chatbot for departments that performs ERP transactions and HR tasks in natural language.
Content Intelligence: Automatic scanning of invoices and contracts → data extraction → summarization, keyword tagging service.
Visual Builder: Real-time detection of inventory status from warehouse footage with AI visual analysis tool.
MLOps for OCI: A fully managed MLOps environment that covers everything from model development to monitoring and governance.
Advantages of introduction
Cost efficiency: High-speed data processing and low-cost operation are achieved by combining with Oracle Exadata and Autonomous Database.
User Experience: Digital Assistant is integrated within your business app and can be used immediately with no training required.
Security: Robust data controls with OCI Identity and Access Management, key management with Vault, and Oracle Audit Vault.
Technology Foundation
OCI integration: Natively works with Oracle Autonomous Database, Oracle Integration Cloud, and Oracle Analytics Cloud.
Edge AI: Lightweight model deployment on Oracle Edge Services supports manufacturing lines where latency is important. |
| 142 | Palantir Foundry | Palantir Technologies Inc. | united states of america | Large-scale data integration/analysis platform. Powerful support for AI model operation and governance. | | | | | https://www.palantir.com/platforms/foundry | DetailsOverview
Palantir Foundry is a platform that delivers large-scale data integration and AI analytics for enterprises. Securely integrate structured and unstructured data inside and outside the company to promote data operations (DataOps). The "Analytics app" built on Foundry creates a workflow from machine learning model training to deployment to evaluation, allowing business users to explore, analyze, and visualize data with drag and drop. It is widely adopted by government agencies, finance, manufacturing, energy, and other industries with high security requirements.
Key Features
Data Integration: Centralize real-time data streaming and batch data management without relying on ETL.
No-touch ML: Equipped with AutoML functionality, you can build, compare, and deploy predictive models without specialized knowledge.
Analytics Apps: Quickly develop dashboards, reports, and ML pipelines with a no-code visual IDE.
Collaboration: Finely control data lineage and access privileges to support cross-team development.
Governance: Provides a framework for account authentication, approval processes, audit logs, and encryption policies.
Advantages of implementation
Agility: Launch data integration projects in weeks that would previously take months.
Improve ROI: Reduce downtime by 30% by introducing an anomaly detection model on the manufacturing line.
Security: A track record of implementing FedRAMP/CJIS compliance for financial institutions and government agencies.
Technology Platform
Cloud & On-Premise: Supports hybrid deployment to AWS, Azure, GCP, as well as your own data center.
Plugin ecosystem: Seamless integration with external ML libraries such as TensorFlow, PyTorch, and Spark MLlib. |
| 143 | PathAI | PathAI, Inc. | united states of america | AI specialized in pathological image analysis. Speed up diagnostic support and biomarker search. | | | | | https://www.pathai.com | DetailsOverview
PathAI is a medical image analysis AI platform designed for pathologists. High-resolution images of tissue section slides are analyzed using deep learning models to automatically detect and quantify cancer cells and other lesions. It ensures the accuracy and reproducibility required for clinical trials and biomarker discovery, and can be incorporated into multicenter joint research and diagnostic support systems. Multiple modules, including FDA 510(k) cleared products, provide powerful support for pathology diagnostic workflows.
Main features
Tissue slide analysis: Automatic segmentation of cancer tissues, inflammatory cells, etc. and mapping at the pixel level.
Biomarker quantification: Quantitative analysis of IHC staining results and integration with gene expression profiles.
Quality control: Detects reading errors and scan defects and automates rescan instructions.
Collaborative review: Specialists share real-time comments and annotations using a web-based viewer.
Research Platform: Provides an integrated analysis pipeline that combines clinical data/genetic data.
Advantages of introduction
Improved diagnostic accuracy: Subjective differences among pathologists are reduced by an average of 20%, making it possible to obtain consistent judgment results.
Improved efficiency: Reduced the time from slide loading to report generation by more than 50% compared to conventional methods.
Research application: Accelerate the discovery of new biomarkers through large-scale analysis of clinical trial cohorts.
Technical platform
Model architecture: Adopts a hybrid network of ResNet encoder + UNet decoder.
Data security: Protect medical data in both HIPAA-compliant cloud environments and on-premises environments.
FDA Approval: Some modules have obtained US FDA 510(k) clearance. |
| 144 | Tempus | Tempus Labs, Inc. | united states of america | Cancer genome data analysis platform. We provide treatment support that combines clinical data and AI. | | | | | https://www.tempus.com | DetailsOverview
Tempus is an AI analysis platform that supports precision medicine with a focus on cancer treatment. We provide "AI-driven decision support" that integrates tumor genomic information (DNA/RNA sequences), electronic health record (EHR) information, and clinical trial data to propose the optimal treatment for each patient. In particular, it is adept at mutation analysis at the molecular level, and uses deep learning using millions of cancer patient data to achieve highly accurate prognosis prediction, treatment response prediction, and clinical trial matching. We are collaborating with research institutions, hospitals, and pharmaceutical companies in the oncology field, and are considered to be one of the companies at the forefront of personalized medicine.
Main features
Tempus xT/xF: NGS (next generation sequencing) test for mutation profiling of solid and blood cancers.
Tempus Lens: Extracts and structures unstructured data such as EHR, radiology images, and clinical notes using natural language processing (NLP).
AI matching: Automatically presents clinical trial candidates and optimal therapeutic drugs based on patient profile.
Tempus One: A smart terminal with a voice interface that enables real-time information call-up during medical treatment.
Biomarker analysis: AI supports prediction of response to immunotherapy and elucidation of resistance mechanisms.
Advantages of introduction
Medical accuracy: Treatment proposals based on patient-specific molecular information improve response rates by up to 30% compared to standard treatments.
Clinical trial efficiency: AI matching reduces screening time by 80% and significantly improves clinical trial registration rate.
Research Support: Accelerate drug discovery for pharmaceutical companies with access to hundreds of thousands of anonymous data.
Technical platform
NLP engine: Equipped with a Transformer-based medical specialized model to enhance the accuracy of information extraction from medical records and pathology reports.
Data security: Compliant with HIPAA, CLIA, and CAP, allowing collaboration with medical institutions.
Cloud infrastructure: Build a secure data integration environment centered on AWS. |
| 145 | Zebra Medical Vision | Zebra Medical Vision Ltd. | Israel | A group of AI models that automatically diagnose medical images. Supports early detection by analyzing X-rays, CT, etc. | | | | | https://www.zebra-med.com | DetailsOverview
Zebra Medical Vision is a company that develops and provides AI solutions specializing in radiological image analysis, and plays the role of an "AI radiologist" that automatically detects disease signs from images such as chest X-rays, CT, and MRI. Equipped with FDA-approved AI models that support over 20 clinical areas such as heart disease, lung disease, bone fractures, cancer, and liver disease, it supports medical institutions' interpretation work. It is particularly effective in screening and preventing oversights, and there are many cases where AI alerts lead to early treatment.
Main features
Chest X-ray analysis (AI1 Chest): Detects over 14 types of lesions including pulmonary nodules, pleural effusions, cardiomegaly, and atelectasis.
Fracture detection: Detects vertebrae compression fractures and proximal femur fractures with high accuracy, and is also used to aid in the diagnosis of osteoporosis.
Cardiac CT analysis: Automatically calculates coronary artery calcium score and quantifies heart disease risk.
Liver disease evaluation: Automatic evaluation of the progression of liver steatosis (NAFLD) and liver cirrhosis using CT/MRI.
Public data collaboration: Continuous learning of AI is carried out in collaboration with Israel's largest medical examination database.
Advantages of introduction
Prevention of misdiagnosis: Second interpretation using AI reduces the rate of missing important findings by 40%.
Work efficiency: Automatic judgment of routine examinations reduces radiologist's interpretation time by 30%.
Correcting medical disparities: It is possible to homogenize the accuracy of image diagnosis even in regional hospitals with few image interpreters.
Technology platform
Deep learning: 3D CNN based on DenseNet architecture achieves highly accurate organ recognition and lesion classification.
Approval status: Many analysis algorithms are FDA 510(k) certified. We also offer CE marked products.
Cloud/on-prem support: Flexible system configuration that can be connected to DICOM-compatible PACS. |
| 146 | Tableau AI (Ask Data) | Tableau Software | united states of america | Tableau's natural language query capabilities. Automatically generate graphs when you ask questions in a chat style. | | | | | https://www.tableau.com | DetailsOverview
Tableau AI (specifically the Ask Data feature) is an AI-based natural language interface integrated into Tableau, the well-known visual analytics tool. Traditional dashboard operations require specialized operations to create graphs and adjust filters, but Ask Data automatically generates graphs and indicators by simply asking questions in natural language. For example, just type in "Show me regional sales in 2023 in a pie chart" and it will be visualized instantly.
This feature provides powerful data analysis capabilities even to general users who have no experience with SQL or BI tools, and is attracting attention as a tool that breaks down the barrier of data literacy.
Main features
Natural language queries (Ask Data): Analyzes natural sentences, mainly English, and generates optimal graphs and charts.
Explain Data: AI estimates and presents statistical factors for abnormal values and trends.
Data Stories: Summarizes important analysis results in text and supports automatic report generation.
Einstein Discovery integration: Links with Salesforce's AI model to enable prediction and trend analysis.
Voice interface (experimental function): In the future, we are also developing a function that allows visualization from voice input.
Advantages of introduction
Data democratization: Analysis can be performed by sales, human resources, and accounting departments other than the data department.
Save time: Graph generation completes in seconds without complex SQL or filter settings.
Narrative analysis: By explaining the analysis in text, it is possible to create reports that are easy to understand for management.
Technical platform
Natural language processing (NLP): Automatically converts queries to suit data structures using a unique parsing engine.
Statistical analysis module: Explain Data automatically executes multivariate regression and pattern recognition.
Salesforce integration: Integration with Einstein Discovery allows predictions from CRM data. |
| 147 | ThoughtSpot | ThoughtSpot, Inc. | united states of america | A BI tool that advocates “Search & AI-driven Analytics.” Interact with dashboards using natural language. | | | | | https://www.thoughtspot.com | DetailsOverview
ThoughtSpot is a BI (Business Intelligence) tool with an innovative concept of "automating dashboards with search and AI." Particularly noteworthy is the idea of ``Search & AI-Driven Analytics,'' which enables search-based analysis that provides insights simply by asking questions in natural language.
The company was founded by engineers from Google, and is characterized by a UX design that specializes in data search. Even if you are not a data analyst, you can simply enter "year-on-year sales growth rate" or "top 10 products" as if searching on Google, and visualization and factor analysis will be performed in real time. We also collaborate with large amounts of data sources (Snowflake, BigQuery, AWS, etc.) to promote improved data utilization across the company.
Main features
Search-based visualization: Instantly generate bar graphs, line graphs, etc. from text searches.
SpotIQ Insights: AI automatically searches and presents trends, abnormal values, and causal relationships.
Liveboards: Interactive dashboard functionality that updates in real time.
ThoughtSpot Everywhere: SDK/API group that allows you to embed ThoughtSpot functionality into your own applications.
Mobile compatible: iOS/Android app supports analysis while on the move.
Advantages of implementation
No-code analysis: Even business departments can extract advanced insights without SQL.
Real-time performance: Connect directly to cloud database and always analyze with the latest data.
Insight discovery support: AI automatically reveals trends that you are not aware of.
Technical platform
Original search engine: Very high accuracy in converting natural language to structured queries.
AI Analyst (SpotIQ): Using unsupervised learning, it also suggests data relationships that the user is unaware of.
Cloud native: Integrates with AWS/Azure/GCP and can be quickly deployed in SaaS format. |
| 148 | Qlik Sense | QlikTech International AB | Sweden (founding place), currently headquartered in the United States | A BI platform with an associative engine and AI recommendation function. | | | | | https://www.qlik.com | DetailsOverview
Qlik Sense is a next-generation BI tool that combines an "Associative Engine" and AI functionality. Unlike traditional fixed dashboards, its greatest strength is that users can follow related data associatively with each click, allowing them to "explore data along the flow of their thoughts." Furthermore, it is equipped with an AI-powered "Insight Advisor" that can automatically present charts and trends that are most appropriate for your questions.
Main features
Associative data engine: Automatically detects connections between all data and enables non-linear exploration.
Insight Advisor: AI presents charts, trends, and factors from natural language input and click operations.
SaaS/on-premise compatible: Can be implemented in both cloud and corporate environments.
Self-service BI: Drag and drop data filtering and axis changes.
Qlik AutoML: Build machine learning models with no code and easily perform predictive analysis.
Advantages of introduction
Good at exploratory analysis: Even without a hypothesis, you can discover related data according to your interests.
Cross-departmental analysis: Integrates marketing, sales, inventory, etc., accelerating decision-making speed.
Ensuring governance: Achieves flexible operation on the user side while linking with the corporate data infrastructure.
Technology Foundation
Qlik Associative Engine: Instantly visualize nonlinear connections between data models.
Natural Language Analysis: Insight Advisor analyzes search queries and presents insights in an interactive format.
ML integration: Advanced analysis algorithms can be incorporated by linking Qlik AutoML and Python/R. |
| 149 | Pi（Personal AI） | Inflection AI | America | A personal AI assistant specialized in dialogue. | | | | | https://pi.ai | DetailsCharacteristics: Good at emotional conversations. Aim for responses that are more mental health-oriented. |
| 150 | Krutrim AI（Koo AI） | Krutrim | India | Conversation AI specialized for the Indian market. Multilingual support (including Hindi). | | | | | https://www.krutrim.com | DetailsFeatures: Collaborates with the chat app "Koo" to strengthen regional language support. |
| 151 | Cognosys | Cognosys AI | India | AI OS that supports multi-agents. A unique structure that allows multiple AIs to operate in parallel. | | | | | https://www.cognosys.ai | DetailsFeatures: Intuitive AI UI without prompts. |
| 152 | Phind（旧名：Hello Cognition） | Phind | America | Search/Q&A specialized AI for developers. | | | | | https://www.phind.com | DetailsFeatures: High-speed search of information on Stack Overflow and GitHub and present optimal code. |
| 153 | Motorola SVX + Assist | Motorola Solutions | America | Public safety, police/fire service support | | | | | 詳細 | DetailsFeatures: A device that integrates a body camera, speaker microphone, and AI assistant "Assist." Reduce reporting work for patrol operations by up to 40%, including automatic input of license plate and driver's license information. Co-developed with 32 US public safety agencies. |
| 154 | Meta AI Age Detection | Meta | America | Age detection and minor protection on Instagram | | | | | 詳細 | DetailsFeatures: Uses AI to estimate user age and automatically enforce privacy settings for accounts of minors. Even if a false positive is detected, the user can restore the settings. |
| 155 | Poe | Quora, Inc. | united states of america | (Quora): A platform where you can try out various AI models (Claude, ChatGPT, Llama, etc.). | | | | | https://poe.com | DetailsPoe is a multi-AI chat platform provided by Quora. The name Poe stands for "Platform for Open Exploration," and its main feature is that users can use and compare multiple AI models across the board. Various LLMs (Large-Scale Language Models) such as ChatGPT (OpenAI), Claude (Anthropic), GPT-4, Gemini (formerly Bard, provided by Google), Mistral, Meta's LLaMA, and even Baichuan from China can be utilized with one interface.
In addition to chatting, Poe also provides the ability for users to build their own AI agents called "bots." This also allows you to set up custom bots for specific purposes and share them with other users. Because Quora has a large amount of Q&A data as a background, it is possible to respond to user questions using a wealth of background knowledge.
In addition, it supports both web apps and smartphone apps (iOS/Android), and the interface is intuitive. It is designed to be easily used globally as it supports multiple languages including Japanese. |
| 156 | Llama | Meta Platforms, Inc. | united states of america | (Meta): An open source large-scale language model series developed by Meta. | | | | | https://ai.meta.com/llama | DetailsLLaMA (Large Language Model Meta AI) is an open source large language model series developed by Meta Platforms (formerly Facebook). LLaMA 2, which appeared in 2023, was developed to compete with OpenAI's GPT and Google's Gemini, and has been widely released for commercial use. The feature of LLaMA is that while it has high-precision natural language processing performance, it is designed to be relatively lightweight and can operate on personal PCs and on-premises environments.
Meta is committed to promoting transparency in AI research through LLaMA research, and is committed to academic contributions and open community building. It is also attracting attention as a base model for companies to develop their own AI applications, and for companies that want to use AI without exposing private information to the outside world.
``LLaMA 3'' is also scheduled to be announced in 2024, and is said to have performance and scalability comparable to GPT-4, and further leaps are expected. It is an open source LLM that is currently accessible through Hugging Face and is extremely popular in the researcher and developer community. |
| 157 | Mistral AI | Mistral AI SAS | France | Model: Featured as a high-performance open source model (e.g. Mistral 7B, Mixtral 8x7B). | | | | | https://mistral.ai | DetailsMistral AI is a French AI startup founded in 2023 and a leader in large-scale language model (LLM) development from Europe. The founders are former researchers at Meta and Google DeepMind, and the company has attracted attention, having successfully raised over $100 million in funding in just six months since its founding.
Mistral's models are completely open source, yet extremely accurate. Although the representative model "Mistral 7B" is small with 7 billion parameters, it has a performance that exceeds LLaMA and GPT-3.5. In addition, "Mixtral 8x7B" is an innovative design that uses the MoE (Mixture of Experts) method to increase output accuracy while reducing calculation costs by utilizing specific expert models as needed.
One of its features is its design in consideration of European data regulations (GDPR, etc.), making it a model LLM that is suitable for the European market, which is strict on privacy. Currently, API provision has started, and implementation is progressing particularly in companies, government agencies, and educational institutions in the EU. |
| 158 | Pika Labs | Pika Labs, Inc. | united states of america | A popular tool for generating videos from text. | | | | Image/video generation and editing: | https://www.pika.art | DetailsPika Labs is an innovative AI video generation startup launched in 2023 that provides a platform that allows users to generate high-quality animations and videos simply by inputting text and images. What attracted particular attention was its ability to ``automatically generate several seconds of video from a prompt,'' and it has become an important feature in promoting the democratization of video production.
This AI generates high-visual-quality videos lasting from a few seconds to tens of seconds simply by giving instructions in natural language such as ``a cat flying in space'' or ``a robot walking in a futuristic city.'' Pika Labs in particular is characterized by its excellent diversity of styles and the ability to handle various visual expressions such as anime style, live-action style, cyberpunk, etc.
The UI is extremely simple and designed to be intuitive even for users who have never used AI, and is used by a wide range of people, from creators to marketers and educators. The generated video can be used as is for SNS posts, advertisements, presentations, game content, YouTube shorts, etc.
In 2024, major functional enhancements will be implemented as "Pika 1.0". It has evolved into a professional video production tool, allowing you to resize videos, direct camera motion, and specify the movement of specific objects. It is currently available as a web-based service, and is also heavily integrated with Discord. |
| 159 | Magnific AI | Magnific AI | united states of america | This is an advanced image upscaling tool. | | | | Image/video generation and editing: | https://magnific.ai | DetailsMagnific AI is an AI tool that specializes in “high-precision remastering” of images and illustrations, and can reconstruct low-resolution images and rough illustrations into surprisingly high-quality, detailed visuals. While conventional "upscaling AI" simply increases the resolution, Magnific uses a unique technology in which AI creatively redraws parts of the image and converts it into an artistic and realistic finish.
Users can obtain professional-looking results by simply uploading the original image and adjusting the desired style and detail strength using the sliders. It is compatible with a variety of genres, including illustrations for anime and games, manga manuscripts, and photo corrections, and is particularly well received by artists and concept designers.
Magnific AI is also attracting attention for its ability to utilize the "room for imagination" provided by AI. By "creatively supplementing" minute details of backgrounds and characters, it is possible to bring out new appeal that the user did not intend. In this respect, it stands as a ``creative AI between image generation and upscaling,'' and differentiates itself from existing AI image tools.
Currently, we have moved from closed beta to open beta and are offering it as a monthly subscription. |
| 160 | Krea AI | Krea Technologies, Inc. | united states of america | A tool that allows you to generate and edit images in real time. | | | | Image/video generation and editing: | https://www.krea.ai | DetailsKrea AI is a next-generation creative tool that specializes in "real-time visual design support using AI." While conventional image generation AI works by inputting text and generating a single image a few seconds later, Krea is revolutionary in that it achieves "real-time visual editing and generation."
When a user sketches or marks up on the canvas using the pen tool or shape tool, the AI instantly converts it into a beautiful image. This can be described as a fusion of image editing software and an AI generation engine, and it brings great benefits especially to professions that require intuitive and quick feedback, such as UX/UI designers, web creators, manga artists, and illustrators.
Krea is uniquely adjusted based on existing image generation models such as Stable Diffusion and ControlNet, and has high controllability such as pose control, style transfer, and layout preservation, and allows highly accurate generation without relying on prompts.
Furthermore, features for creators such as collaboration with Photoshop plug-ins, PNG output, and maintaining transparent backgrounds have been implemented, and expansion to 3D data and animation is planned in the future. |
| 161 | Udio | Uncharted Labs | united states of america | Like Suno AI, it is a tool that can generate high-quality music. | | | | Music generation: | https://www.udio.com | DetailsUdio is an innovative AI tool that generates music from text, and is attracting attention for significantly lowering the barrier to music production. The developer, Uncharted Labs, is a startup founded by former Google DeepMind researchers and is developing cutting-edge products that combine generative AI and voice processing technology.
The biggest feature of Udio is that users can automatically generate high-quality music in seconds to tens of seconds by simply entering a simple text prompt. By specifying the genre and atmosphere (e.g., ``calm jazz-style instrumental song'', ``up-tempo EDM-style track,'' etc.), you can easily generate original sound sources even if you are not a professional musician, and it is also used for BGM of YouTube videos, advertisements, podcast production, etc.
Technically, Udio uses a Transformer-based deep learning model that learns patterns, structures, and genre-specific rhythms from large amounts of music data. In addition, a workflow has been built that performs lyric generation (lyrics) and melody generation separately, and automatically processes everything up to voice synthesis and mixing. This allows anyone to create high-quality sound sources on demand without copyright issues.
Currently, Udio offers free and paid plans for individual creators, and is gradually gaining its presence in the music industry and video production market as it allows for the production of professional sound sources in a short time. A license plan based on API collaboration and commercial use is also being planned, and it is expected that it will be expanded to companies, educational institutions, and the entertainment industry in the future. |
| 162 | Blackbox AI | Blackbox AI | united states of america | It is an AI tool specialized in code generation and search. | | | | For developers: | https://www.useblackbox.io | DetailsBlackbox AI is a code completion and search assistance tool designed primarily for developers, and is an AI coding assistant that is often used in conjunction with GitHub. The developer, Blackbox AI, aims to streamline "knowledge search" and "automatic code generation" in the software development process, and is rapidly gaining popularity among startups and individual developers.
This tool has similar functionality to ChatGPT and Copilot, but is especially strong for reference-oriented tasks such as ``searching existing codebases'' and ``reusing code.'' Its ability to cross-analyze GitHub repositories and suggest relevant code snippets for user queries (e.g. "How to validate forms with React") is highly valued.
Blackbox AI's distinctive features include:
✅ Fast code snippet search: Natural language-based search in over 50 languages
✅ Code completion: Works with editors (such as VSCode) to provide real-time completion
✅ Screen code capture: AI reads the code displayed on the screen, recognizes it as text, and makes it reusable
✅ Automatic comment generation: Insert appropriate comments and explanations for existing code
Basic code search and completion functions are available on the free plan, but paid plans enable expanded GitHub integration, team collaboration functions, and advanced search operations. Blackbox AI is also designed with privacy in mind, allowing analysis of the local codebase with the user's permission.
As an AI tool that streamlines development and learning, many engineers have incorporated it into their daily coding workflows, and it has come to be known as "your own personal Stack Overflow." |
| 163 | FigJam AI / Miro AI | Figma, Inc. | united states of america | AI features within online whiteboard tools (e.g. automatic generation of diagrams). | | | | Other/specific uses: | https://www.figma.com/figjam | DetailsFigJam AI and Miro AI are both generative AI capabilities built into online whiteboarding tools that help teams brainstorm, workshop, flowchart design, UX planning, and more more efficiently and creatively. Although these are products of similar genres, each has its own individuality based on the philosophy and design philosophy of each developer company.
🔶 FigJam AI (Figma, Inc.)
FigJam is a sister product of the design tool "Figma" and is a lightweight collaborative whiteboard tool. FigJam AI is an extension that allows you to automatically generate diagrams, idea maps, UX flows, sticky notes, and more from text-based prompts.
Key features:
🧠 Automatic idea development: Users just type a few words to automatically generate mappings and sticky notes for related ideas
🔄 Meeting note summaries: Summarize comments written in real time and organize them into action items and takeaways
🔧 Automatic template creation: Instantly build board structures for custom workshops and UX design
👥 Collaboration support: AI automatically suggests adjustments and classifications while multiple people are editing simultaneously
Seamless integration with Figma products makes it an excellent tool to bridge the gap between design teams and the business side.
🔷 Miro AI (Miro)
Miro is a whiteboard platform provided by a global company originating from Europe, and is positioned similar to FigJam, but it has many specialized functions for larger projects and companies. Miro AI is an intelligent support system that automates everything from meeting preparation to mind mapping, ideation, and analysis tasks.
Featured features:
🔍 Automatic summarization and insight extraction: Concisely summarize the contents of team discussions and incorporate them into the next action
🪄 Mind map generation: Expand related ideas radially from one keyword
📊 Diagramming support: Generate diagrams, timelines, and business flows from text
🧩 Collaboration with other AI tools: Advanced interlocking is also possible by collaborating with ChatGPT and Notion AI.
Miro AI is increasingly being introduced for enterprise use, and is used in a variety of industries such as product development, education, consulting, and marketing. |
| 164 | ClickUp AI / Asana Intelligence | Mango Technologies, Inc. (ClickUp) | united states of america | AI capabilities within project management tools. | | | | Other/specific uses: | https://clickup.com | DetailsClickUp AI and Asana Intelligence are both generative AI capabilities built into business and project management platforms that automate tasks, generate documents, support decision-making, and more. What they all have in common is that they dramatically streamline daily business processes, and their value increases, especially as your team grows.
🔶 ClickUp AI (Mango Technologies, Inc.)
ClickUp is an "all-in-one" business efficiency tool that integrates project management, task management, document sharing, time tracking, and more in one platform. ClickUp AI is an AI feature built into this core that acts as a powerful assistant to accelerate project operations.
Key features:
🧾 Automatic generation of task summaries and descriptions: Automatic generation of easy-to-understand task descriptions and checklists from prompts
📝 AI generation of meeting notes and reports: Automatically organize meeting content into structured reports
🔍 Custom prompt support: Templates and workflow suggestions optimized for your team
📚 Knowledge management support: Extract important insights from documents and convert them into tasks
ClickUp AI is truly designed as "AI for work," with "Role-Specific AI Packs (Marketer, Developer, Sales, etc.)" that allow you to select specialized prompts tailored to your industry.
Asana Intelligence (Asana)
Asana is a SaaS that specializes in project management and workflow optimization and is deployed by many global companies. Asana Intelligence is a generative AI feature based on OpenAI technology that visualizes work and speeds up decision-making.
Notable features:
📊 Project summary/progress analysis: Concisely summarize the progress status of multiple tasks and projects
📌 Task decomposition/renaming support: Convert ambiguous request details into clear tasks
🧠 Insight presentation: Infer and present "what is the bottleneck" from past trends
✏️ AI draft creation: Generate documents and update reports based on conversations
Asana Intelligence is strong in understanding the context of work and has the intelligence to assist human judgment. In the future, it is expected to be linked at the AI level with chat tools such as Slack, and to automate tasks and reminder suggestions. |
| 165 | Zoom AI Companion / Google Meet (Gemini連携) / Microsoft Teams (Copilot連携) | Zoom Video Communications, Inc. | united states of america | AI functions such as summarization and transcription within the web conferencing tool. | | | | Other/specific uses: | https://explore.zoom.us/en/products/ai-companion | DetailsThese three are generative AI assistants built into their respective video conferencing/collaboration platforms, and are attracting a lot of attention for improving work productivity in remote and hybrid work environments. It is equipped with advanced AI functions that can extract important points during meetings, create minutes, organize action items, and even perform advance preparation and follow-up.
🔷 Zoom AI Companion (Zoom Video Communications, Inc.)
Zoom AI Companion is Zoom's unique generation AI function that works in conjunction with Zoom's conference, chat, email, whiteboard, and other functions. We will combine OpenAI technology and Zoom's own AI to improve productivity in real time.
Main features:
✍️ Automatic summarization of meetings: Summarizes what was said in real time and saves it as minutes
✅ Extracting action items: Automatically organizes and visualizes "things to do" that came up during the meeting
🔍 Chat summary and main points display: Helps understanding by condensing long chat history into a few lines
📧 Email draft support: AI creates replies and notifications for Zoom Mail and other services
Zoom AI Companion is designed with privacy in mind, and user data is not used to train the AI model.
Official website: https://zoom.us/ai-companion
🔷 Google Meet (integrated with Gemini)
Google Meet is gradually providing AI support functions in conjunction with Google's large-scale language model "Gemini (formerly known as Bard)". A particular strength is the seamless integration with Google Workspace (Gmail, Docs, Calendar, etc.).
Featured Features:
🗒️ Real-time meeting memo generation: Automatically organizes discussion content and outputs to Google Docs
🧠 Background information presentation: Automatically extracts context from meeting participants' schedules and documents
🔁 Summary for absentees: Automatic notification of summaries and actions for those who missed the meeting
🔍 Voice command operation: A function to ask AI questions and get answers from meeting records is also planned to be implemented.
By linking with the entire Google Workspace, it will be dramatically easier to organize information on a project-by-project basis.
Official website: https://workspace.google.com/intl/ja/meet/
🔷 Microsoft Teams (Copilot integration)
Microsoft Teams provides AI support for Teams meetings by integrating Microsoft's "Copilot" technology. Integration with the entire Microsoft 365 (Outlook, Word, Excel, PowerPoint, etc.) is extremely powerful.
Notable features:
🧾 Generation of minutes and transcripts: Summarize conversations in natural sentences and extract key points
📌 Presentation of action items and proposals: Clearly record who should do what
📊 Preparation materials generation before meetings: Provides a summary of related materials, and can also create materials based on what was said
📈 Document collaboration: Automatically generate PowerPoint and Word drafts based on the meeting content
In particular, "Copilot in Teams" is designed to support AI in all phases before, during, and after the meeting, dramatically improving the way you use your time.
Official website: https://www.microsoft.com/ja-jp/microsoft-teams/copilot |
| 166 | Glean | Glean Technologies, Inc. | united states of america | AI search platform for businesses. | | | | Other/specific uses: | https://www.glean.com | DetailsGlean is a generative AI tool for enterprises that was developed to be like a "Google search" for internal knowledge. Glean connects all your internal information sources (Google Drive, Slack, Confluence, Notion, Jira, Outlook, etc.) and lets you search and summarize the information you need in natural language. It is an innovative tool that aims to solve the inefficiency caused by knowledge fragmentation and fundamentally changes the problem of not being able to find the information you want.
🔍 Key Features and Functions
🔗 Integrates with hundreds of SaaS
Glean can integrate with major business tools such as Google Workspace, Microsoft 365, Dropbox, Salesforce, Zendesk, Notion, and more. Scattered information can be searched all at once.
💬 Natural Language Search and Generated Answers
For natural language questions such as "What are the latest meeting notes for project 〇〇?", Glean finds the relevant document or thread and responds with a summary of the key points.
🧠 Automatically map organizational knowledge
AI learns user behavior and team composition and presents information such as ``Who is familiar with which field'' and ``Who should I ask about this matter?'' We systematize knowledge sharing, which tends to be individualized.
✅ Permission-based access control
Information search results are restricted according to each user's viewing privileges, preventing the risk of leaking confidential information.
📊 Dashboards and Analytics
Reports that visualize frequently searched keywords and knowledge gaps within your organization help improve information sharing.
📈 Glean Value and Adoption Benefits
Glean significantly reduces information search time, increasing employee productivity and facilitating onboarding and cross-team collaboration. In addition, the high accuracy and speed of answers provided by AI enables knowledge sharing to be more immediate than traditional "internal wikis" and "FAQs."
Especially for global companies, startups, and distributed teams, "who knows what" and "where information is located" determine operational efficiency, so tools like Glean have become important at an infrastructure level. |
