We don't have a tool for that in the directory yet. Try a different search term.
Chatbots & Assistants
30OpenAI's best-known AI chatbot for text, questions, images, files and voice conversations.
Anthropic's AI assistant, strong at writing, analyzing long documents and coding.
Free chatbot from Chinese company DeepSeek with strong reasoning and math abilities.
Google's AI assistant connected to Google Search, Gmail, Docs and other Google services.
Moonshot AI's chatbot, known for long context, web search and agentic tasks.
Apple's voice assistant, from iOS 27 a standalone app with on-screen awareness and camera mode.
Meituan's chat and API featuring the multimodal agent model LongCat-2.5 with a 1-million-token context.
Meta's AI assistant, available on the web and in WhatsApp, Instagram and Facebook.
Quora's platform for using many AI models from different providers in one app.
Open-source chat app for power users with model comparison, personas and voice.
AI assistant built into the Brave browser that summarizes pages and answers questions.
Desktop and mobile app to chat with various AI models via API key or locally.
Open-source desktop app bundling many AI providers and local models in one interface.
DuckDuckGo's anonymous AI chat offering several models without an account or tracking.
Baidu's chatbot based on ERNIE models, mainly for Chinese users.
Chatbot from French company Mistral AI with web search, document analysis and images.
Open-source ChatGPT-style chat interface combining many AI providers and self-hosting.
Modern open-source chat app with plugins, agents and support for many models.
Privacy-focused AI assistant from Proton with encrypted chat history.
Microsoft's AI assistant, integrated into Windows, Edge and Microsoft 365.
Lightweight open-source chat interface that can be self-hosted in a few clicks.
Free Alibaba chatbot using Qwen models, including image and video generation.
AI companion app for personal conversations with a virtual friend.
Alternative chat interface that uses multiple AI models with your own API keys.
Private, uncensored AI chat using open models that doesn't store conversations on servers.
Zhipu AI's chatbot using GLM models, also for creating slides and web pages.
Coding
34The most popular open-source coding agent: it plans, writes, and tests code right inside your project.
An AI pair programmer in your terminal that writes code right in your project.
Anthropic's coding agent for terminal and IDE that reads, writes and tests code and runs commands.
AI autocomplete and chat right inside VS Code – with local models.
OpenAI's coding agent as terminal tool, IDE extension and cloud service, usable with a ChatGPT plan.
AI code editor based on VS Code with chat, autocomplete and agents that edit entire projects.
Google's AI development environment where agents write, test and verify code in the browser.
App for running several coding agents in parallel in separate Git worktrees.
Cognition's autonomous AI software engineer that plans and implements tasks on its own.
Very lightweight open-source coding agent that solves software tasks using only shell commands.
Alibaba's AI coding environment with agents that write code and handle tasks in a project.
Claude Code skill that builds interactive 3D machine models as a single offline HTML file.
Sourcegraph's coding agent for terminal and editor, built for large tasks.
AI coding assistant and agent that understands large codebases particularly well.
In-browser app builder that creates and runs full-stack web apps from prompts.
Open-source coding agent for VS Code that edits files and runs commands with your approval.
Automatically reviews pull requests with AI and leaves review comments.
Google's open-source terminal agent that codes and handles tasks using Gemini models.
GitHub's AI assistant for code suggestions, chat and agents directly in the editor and on GitHub.
JetBrains coding agent for IntelliJ, PyCharm and more that solves tasks in your project.
Google's asynchronous coding agent that completes tasks in GitHub repos and proposes changes.
Open-source coding agent for VS Code and JetBrains with access to many AI models.
Amazon's AI development environment that turns requirements into specs and code.
Builds complete web apps from text descriptions in the browser, no coding skills needed.
AI platform for code reviews, tests and code quality in the editor and pull requests.
Online development environment with an AI agent that builds and deploys apps from descriptions.
Open-source coding agent for VS Code with modes like architect, code and debug.
AI code assistant focused on privacy, can also run on your own servers.
ByteDance's AI code editor with chat and agent mode for whole projects.
Vercel's app and UI generator that builds React interfaces and web apps from text.
Modern terminal with built-in AI agents that run commands and write code.
AI code editor with the Cascade agent that edits multiple files on its own.
Agents & Automation
32A personal AI agent on your own machine that you control through chat apps and that gets tasks done for you
Open-source AI agent system that carries out penetration tests largely autonomously.
CLI tool letting coding agents control native apps via the operating system's accessibility tree.
Open-source agent framework from ByteDance that splits tasks and produces reports, slides or web apps.
Terminal agent that builds described structures in Minecraft Java, with ghost preview and undo.
Perplexity's macOS AI agent that performs tasks on the computer, now also with local models.
Always-on AI agents from Perplexity that handle recurring tasks and communicate via Slack.
Open Zapier alternative for automations with AI building blocks, self-hostable.
Personal agent framework that uses tools and code inside its own Linux environment.
One of the first autonomous agent projects, now a platform for building continuously running AI agents.
Browser extension that automates recurring web tasks such as data scraping with AI.
Library and cloud service that lets AI agents operate websites in a browser on their own.
Agent mode in ChatGPT that completes web tasks using its own virtual browser.
OpenAI's web browser with built-in ChatGPT and an agent mode for web tasks.
Framework and platform where multiple role-based AI agents solve tasks together.
Platform to build AI apps, agents and workflows visually, also self-hostable.
Drag-and-drop tool for assembling AI agents and chatbots based on language models.
AI agent platform that automatically creates research, slides and sheets or handles calls.
No-code builder for AI-powered automations and agents with many app integrations.
Visual editor for building AI agents and RAG workflows, backed by Python.
Builds AI assistants for email, scheduling, sales and support without coding.
Visual automation platform that links apps and integrates AI agents into workflows.
Autonomous AI agent that independently handles research, websites and multi-step tasks.
Workflow automation with a visual editor, AI agent nodes and a self-hosting option.
Open platform for AI agents that write software, run commands and browse the web.
Perplexity's AI web browser with an assistant that summarizes pages and completes web tasks.
Developer-friendly platform for workflows and AI agents with many API integrations.
Platform for building AI agent teams for sales, support and business processes.
Automates browser workflows like forms or downloads using AI and computer vision.
ByteDance desktop app in which an AI agent operates computer and browser via screen control.
Open platform for scripts, workflows and internal tools, including AI steps.
Connects thousands of web apps into automated workflows, including AI agents and AI steps.
Research & Search
27Planet's chat assistant for searching a decade-plus satellite imagery archive in plain language.
Web search inside ChatGPT that summarizes current results with sources; plus Deep Research.
Shows a graph of similar works for a paper to help explore a research field.
AI search engine for research that answers questions based on scientific studies.
AI assistant for scientific literature: finds papers, summarizes them and extracts data.
AI search engine and API that finds web content by meaning, built for apps and agents.
Multilingual AI search engine giving sourced answers and turning them into slides or mind maps.
AI mode of Google Search that answers complex questions conversationally with Gemini and web links.
Open-source agent that autonomously researches the web and writes detailed sourced reports.
Ad-free paid search engine with AI summaries and Kagi Assistant for many models.
Open-source AI assistant that searches your own documents and the web, locally or in the cloud.
AI research assistant focused on reliable and academic sources with citations.
Open-source AI search framework that splits questions into sub-searches and merges web answers.
Open-source AI answer engine with a generative UI, self-hostable or usable online.
Open-source deep research agent from LangChain that works with various models and search APIs.
Open-source Perplexity alternative, self-hostable and usable with local models.
AI search engine that answers questions with cited web sources, including a deep research mode.
Tool for discovering related papers and authors via visual citation networks.
AI platform to find, read and explain scientific papers and create literature reviews.
Research tool showing how papers are cited—supporting or contrasting—with an AI assistant.
Free self-hostable metasearch engine, often used as a privacy-friendly search backend for AI tools.
Free AI-powered search engine for scientific papers from the Allen Institute for AI.
Open-source Stanford system that researches the web and writes Wikipedia-style reports with sources.
Search API built for AI agents that returns web results prepared for language models.
AI agent that thoroughly searches scientific literature to find relevant papers for complex questions.
AI search and research assistant with sources, now strongly focused on businesses and search APIs.
Writing & Translation
31Detector that checks whether texts or images were likely generated by AI.
AI platform for marketing and sales copy and automated content workflows.
AI translator known for natural-sounding translations, plus DeepL Write for improving texts.
Duden's spelling, grammar and style checker for German texts, with AI rephrasing.
Website translation built into Firefox that runs locally in the browser instead of the cloud.
Google's free translation service for 200+ languages, including photos, conversations and documents.
Writing assistant that checks grammar, style and tone and rewrites text with AI, mainly for English.
Editor that highlights hard-to-read sentences and passive voice to make writing clearer.
AI platform for marketing copy with brand voice, templates and team features.
Kagi's AI-powered translator that can also adjust tone and style of translations.
Grammar and spell checker for 30+ languages including German, with AI rephrasing; core is open source.
Minimalist online text editor with a built-in AI assistant for feedback and ideas.
Free, self-hostable translation API and web interface with no cloud dependency.
Microsoft's translation service, available via Bing on the web and built into Edge and Office.
German AI writer for marketing content focused on German-language texts and brand voice.
AI service for writing stories and interactive fiction, including an image generator.
Novel-writing environment with a story wiki and support for any AI model.
Naver's translator, especially strong for Korean, Japanese and Chinese.
AI writing assistant for academic papers with language editing and submission checks.
Style and grammar checker with detailed reports, popular with authors of long-form writing.
Tool for paraphrasing, summarizing and grammar-checking texts.
Translator and dictionary with in-context example sentences, plus rephrasing and grammar check.
Affordable AI writing assistant with templates for emails, posts and ad copy.
AI writing tool for fiction authors: ideas, descriptions, scenes and revision.
Grammar and style checker for academic and technical English writing.
Writing aid that uses AI to rephrase, shorten, expand or change the tone of sentences.
Language checking and AI tools specifically for academic English writing.
Enterprise AI writing platform with its own models, style guides and brand rules.
AI writer for blog posts, ad copy and SEO content.
Documents & Productivity
30Turn PDFs, Word files and PowerPoint decks into clean Markdown – tables included.
Elegant document and note app with an AI assistant for writing, summarizing and translating.
Tiny OCR tool by Datalab that converts PDFs and scans to Markdown, text or JSON without a GPU.
AI inside the Notion notes and wiki app: writes, summarizes, searches knowledge and takes meeting notes.
Open-source alternative to Notion and Miro with docs, whiteboard and an AI assistant.
Open-source Notion alternative for notes, wikis and projects, also with local AI models.
Presentation software that creates slides with AI and keeps the layout tidy automatically.
Upload a PDF and ask questions about it in a chat, with references to the relevant passages.
AI in Coda, a mix of doc, spreadsheet and app, for writing, analysis and automations.
Records video calls and automatically delivers transcript, summary and to-dos.
Meeting assistant that records, transcribes, summarizes and makes conversations searchable.
Creates finished presentations, documents and simple web pages from a short prompt.
Gemini AI built into Docs, Sheets, Slides, Gmail and Meet for writing, analyzing and summarizing.
AI notepad for meetings that listens without a bot and turns your notes into clean minutes.
Answers questions about long PDFs and documents and summarizes them with citations.
Analyzes spreadsheets and data files via chat and creates charts and reports.
Open-source meeting assistant that transcribes and summarizes conversations fully locally.
AI note app that automatically organizes and links notes and answers questions about your knowledge.
Automatically turns text into diagrams, infographics and visuals for presentations.
Google tool that summarizes your documents, PDFs and links, answers questions and creates audio overviews.
Self-hosted open-source NotebookLM alternative for researching your own sources and generating podcasts.
Transcribes meetings live and creates summaries and action items for Zoom, Teams and Meet.
Collaborative presentation app with an AI generator for slide drafts and text.
Add-on for Google Slides and PowerPoint that creates and edits presentations with AI.
Open-source presentation generator that creates slides with AI, also locally, with PowerPoint export.
Summarizes meetings, emails and chats and provides minutes and conversation analytics.
Encrypted note app with backlinks, an AI assistant and voice transcription.
Spreadsheet with a built-in AI analyst that analyzes and summarizes data and writes formulas.
Browser extension that transcribes and summarizes Google Meet, Zoom and Teams meetings live.
Meeting recorder for Zoom, Meet and Teams with AI notes, summaries and CRM integration.
Image Generation
45The pro tool for image and video AI – with ready-made templates.
Images from text in under a second – perfect for a first taste in Colab.
LoRA adapter for ComfyUI that carries an anime character from a single reference image into new images.
A new adapter for the anime image model Anima picks up a character's look, and partly their style, from a single reference image.
App for Mac, iPhone and iPad that runs image models like Stable Diffusion and FLUX offline.
BRIA AI image models for generating and editing images, free for non-commercial use only.
Comfy-Org has repackaged BRIA AI's Fibo 2.0 (dev) and Fibo 2.0 edit-turbo image models for ComfyUI.
Image model family by Black Forest Labs; some variants freely downloadable, Pro versions via API.
Black Forest Labs has published the model flux-3-image-klein on Hugging Face.
Image and video generator by xAI, integrated into the Grok app and X.
Krea 2 fine-tune that places comic panels, speech bubbles and text precisely via coordinate boxes.
Open image generator based on MOSS-Image, with turbo LoRAs for images in just 8 steps.
Open 2B-parameter text-to-image model by Linum that works directly in pixel space.
Linum, a tiny research lab, trained its own text-to-image model from scratch for about $21,000 in compute.
Open text-to-image model by OpenMOSS that places text and objects precisely via JSON layouts.
The OpenMOSS team has released MOSS-Image-SFT, a fine-tuned text-to-image model with 5.1 billion parameters under the Apache 2.0 license.
NVIDIA's PiD replaces the VAE decoder in image generators like FLUX and SDXL.
Compressed, faster version of the MiniMax-H3 image model by Pruna AI, generating and editing images in 8 steps.
Open image model by Alibaba, strong at text in images (including Chinese) and image editing.
Alibaba has released Qwen-Image-2.1, an image model for generating and editing pictures that also runs locally.
Adobe's image generator, trained on licensed data and integrated into Photoshop and other Adobe apps.
Classic browser interface for generating images locally with Stable Diffusion and extensions.
Free image generator by Microsoft, usable in the browser and the Bing or Copilot app.
Image generation and editing directly inside ChatGPT, good at text in images and precise instructions.
Large community platform for image models and LoRAs, with a built-in online generator.
ByteDance creative platform featuring the Seedream image model for high-quality images and editing.
Simple local image generator interface that delivers good results without many settings.
Freepik's image generator that bundles several top models in one interface.
Image generator especially good at rendering legible text in images, e.g. for posters and logos.
Local creative interface with canvas, layers and inpainting for Stable Diffusion and FLUX.
Creative platform with real-time image generation, many models, upscaling and video.
Web platform for generating images with many styles and models, popular for game art and design.
Well-known image generator that creates artistic, high-quality images from text descriptions.
Google's image model in the Gemini app for generating and precisely editing images via text.
Beginner-friendly image generator with many models and community challenges.
Image generator for designers that also creates vector graphics, icons and consistent brand styles.
Image generator with strong prompt adherence and good typography, including text-based editing.
Widely used open image model by Stability AI, runs locally and has a huge ecosystem of add-ons.
Faster, memory-efficient fork of the AUTOMATIC1111 interface, also supporting FLUX.
User-friendly interface for local image models that uses ComfyUI under the hood.
Online platform for generating images with thousands of community models and LoRAs.
Image Editing & Design
35Open model from vivo that edits images via text instructions natively in 4K, using only 4 inference steps.
AI model in Canva that edits existing designs from text instructions, with results staying editable.
ComfyUI node that cuts out people with a click, with hair-level edge detail up to 4K.
Open 3D model that refines coarse meshes into high-resolution ones using a reference image.
Free Apple model that turns a single photo into a rotatable 3D scene (Gaussian splats).
Blender add-on that uses ComfyUI to create textured, rigged 3D models from text.
Professional image editor with AI features like Generative Fill and Expand powered by Adobe Firefly.
High-precision open model for cutting out objects, including fine hair and edges.
Online design tool for graphics and social posts with AI features (Magic Studio).
Node-based desktop app for image processing and upscaling pipelines with AI models.
Web toolkit to remove backgrounds and objects, upscale and relight images.
Collaborative design tool for interfaces and graphics with built-in AI features.
Open 3D generation model from Tencent that creates textured 3D objects from images.
Online design tool for logos, typography and merch designs with AI assistance.
Plugin for the Krita paint app using Stable Diffusion/FLUX for painting, inpainting and outpainting.
Online service for upscaling and enhancing photos with AI.
AI logo maker that creates logos and brand kits for businesses.
Photo editor with AI tools for skies, skin, lighting and object removal.
AI upscaler that adds new detail while enlarging images; part of Freepik.
Generates textured 3D models from text or images.
Free Photoshop-like image editor in the browser, including AI features.
App for product photos: cut out and replace backgrounds and enhance images with AI.
Photo and design app with AI effects, background removal and image generation.
App for product images with AI background removal, object removal and upscaling.
Browser image editor with AI tools like background removal, generative fill and upscaling.
Web service that automatically removes the background from photos.
Desktop software for denoising, sharpening and upscaling photos with AI.
AI service that quickly creates 3D models from images or text.
Video
51Open-source app from Lightricks: generate videos with sound locally using LTX-2.5 and edit them right in the built-in editor.
Fast open video model by Lightricks that also runs on consumer GPUs.
Generate videos from text or an image – one of the best open video models.
Paris AI lab Kyutai has released CADRE, a family of seven video autoencoders for interactive world models.
Nvidia has released four new models in its Cosmos 3 family with freely downloadable weights.
ByteDance/CapCut creative platform with the Seedance video model and image generation.
Nvidia has released Flora, an open data engine that captions and scores robot videos.
Google's video tool for creating and editing videos, including AI-generated clips.
ComfyUI nodes that create multi-minute 1080p videos with MiniMax H3 by chaining segments.
AI video platform with camera moves, effects and access to multiple models.
Tencent Hunyuan has released Hy-Nomas, an open-source model that builds interactive video worlds piece by piece.
Compressed video model based on MiniMax-H3 that generates short clips with stereo audio and runs locally as GGUF.
Free desktop app that drives an existing ComfyUI and turns an idea into an AI short film in six steps.
Kandinsky Lab has released Kandinsky 6.0 Video: two models (29 and 3 billion parameters) that generate video complete with dialogue, …
Kandinsky Lab has released Kandinsky S2V 1.0, a model that takes a single image and a voice recording and generates a video of that …
LTX, a Lightricks spin-off, has released LTX-2.5, a freely downloadable 22-billion-parameter video model.
Open model from Nvidia and MIT that regenerates low-res MiniMax H3 videos at 2K up to 4K.
Local ComfyUI-based studio that generates videos, images and 3D objects on your own GPU.
Free web app that locally replaces the main person in a video with a character from a photo.
Open video model from Pika that generates videos with synced stereo audio in few sampling steps.
Framework for creating and rendering videos with React code; larger companies need a paid license.
Community ComfyUI project that re-renders videos of people from new camera angles.
Open 27B model that writes, renders and revises explainer and motion-design videos as code.
Open model that swaps people in videos and edits scenes, with an MCP server for agents.
The Vimoo-AI team has released the weights of its video model Vimoo-1.0-Lite for ComfyUI.
Popular video editor with AI captions, effects and automatic editing.
Makes photos talk and creates videos with digital avatars.
Video and podcast editor where you edit via the transcript, with AI helpers.
MiniMax's video generator with strong motion quality and prompt adherence.
Generates expressive talking characters from an image and audio.
Creates videos with talking AI avatars and translates videos with lip sync.
Tencent's open video model for high-quality text- and image-to-video generation.
Creates complete videos with scenes, voiceover and music from a text prompt.
Browser video editor with AI subtitles, translation and clip creation.
Kuaishou's video generator, known for smooth motion and longer clips.
Animates portrait photos with expressions and head motion from a driving video.
Luma AI's video generator that turns text or images into cinematic clips.
Automatically cuts long videos into short social clips with captions.
Easy AI video generator with creative effects for short clips.
Creative platform with AI video generation (Gen models) and many editing tools.
OpenAI's video generator that creates realistic clips with sound from text or images.
Creates training and business videos with AI avatars from text in many languages.
Desktop app for AI video upscaling, denoising and frame interpolation.
Online video editor with AI subtitles, translation and avatars.
Google DeepMind's video model, available in Gemini and Flow, generating clips with audio.
Shengshu's video generator for text-, image- and reference-to-video.
Speech & Audio
58Fast, natural-sounding text-to-speech voices – including German, no graphics card needed.
Expressive voices in 23 languages – and voice cloning from just a few seconds of audio.
Speech recognition with Whisper – up to four times faster, for your own scripts.
Dictate instead of typing – press a key, speak, and the text lands in any app.
A dictation app for every platform: press a shortcut, speak, and your text appears right in the app you're using.
A local AI voice studio: clone voices, have text read aloud and dictate – no cloud required
A local ElevenLabs alternative: clone voices, create audiobooks and dub videos, all on your own machine
ServiceNow has released Apriel Voice 27B Preview, an open model that understands spoken language directly, reasons about it, and calls …
Startup Edge0 has released Audio8-TTS-Realistic, an open speech model with just 0.6 billion parameters.
Real-time voice agents from ElevenLabs supporting 90+ languages and thousands of voices.
Highly natural multilingual text-to-speech with voice cloning, dubbing and transcription.
IBM has released Granite Speech 5.0 2B, an open speech model with 2 billion parameters under Apache 2.0.
Open Tencent model that live-translates spoken Chinese and English with about three seconds of delay.
Tencent has released Hy-SimulMT, an open model that translates spoken Chinese and English live, while the person is still talking.
Text-to-speech model with voice cloning that runs in real time on a regular CPU.
KittenML has released code and weights for Kitten-TTS 2, a text-to-speech model with 1.7 billion parameters and ternary weights.
Three new models from Liquid AI's LFM3 family have appeared in the Ollama library: Audio, Omni and Vision-Language, each around 2 …
LinnetTTS is an English text-to-speech model with just 41 million parameters.
Small speech model based on Qwen3-TTS with voice cloning, laugh tags and a fast Turbo variant.
A developer used reinforcement learning and distillation to turn Qwen3-TTS (0.6B parameters) into Lure and Lure Turbo.
Open text-to-speech model for eleven Indian languages and English, with voice cloning from a few seconds of audio.
Ant's inclusionAI team has released the Ming-3.0 speech family under the MIT license.
Mistral model family for speech recognition, text-to-speech and voice agents in 50+ languages.
Nvidia has released Nemotron 3.5 Transcribe, an open speech model you steer with text prompts.
Nvidia has released NemotronLabs RE-USE, a model that removes noise, reverb, clipping and compression damage from speech recordings.
OmniVoice is a free text-to-speech model that can recreate a voice from a few seconds of audio and supports more than 600 languages.
Very small text-to-speech models by Gradium that turn text into speech locally.
Gradium has released its Phonon text-to-speech family in the Ollama library.
Small text-to-speech model from Kyutai that runs on ordinary CPUs, including a German model.
A solo developer spent almost a year building Scylla's Band v4, a free text-to-speech model with only about 131 million parameters.
Very fast, compact text-to-speech model for on-device use; version 3 supports 31 languages.
US startup Underdog has unveiled Woof, a text-to-speech model that weighs in at about 32 MB and works without an internet connection.
Xiaomi has released CocktailASR-1, an open speech recognition model.
Web tool that uses AI to remove echo and noise from voice recordings for studio-like sound.
API for transcription with speaker labels, summaries and real-time speech recognition.
Very fast, realistic AI voices via API, designed for real-time voice applications.
Open speech synthesis with multilingual voice cloning, maintained by the community.
Cloud API for fast speech-to-text and text-to-speech, often used for voice agents.
Open model by Nari Labs generating realistic multi-speaker dialogue, including laughter.
Open audio model by Boson AI for expressive speech generation and voice cloning.
Voice AI with emotionally expressive voices and detection of emotion in speech.
Small, fast open text-to-speech model that sounds good even on modest hardware.
Removes background noise and echo live during calls and video meetings.
Mac app that transcribes audio and video locally with Whisper without uploading data.
Web studio for AI voiceovers for videos, presentations and e-learning.
Open, very fast and accurate speech recognition models from NVIDIA.
Open library for speaker diarization: finds out who speaks when in a recording.
Voice cloning and text-to-speech for businesses, including deepfake audio detection.
Open conversational speech model by Sesame for natural-sounding dialogue voices.
Reads texts, PDFs and web pages aloud with natural AI voices, as app and browser extension.
Open Microsoft model that generates long podcast-style conversations with multiple voices.
OpenAI's open speech recognition model that transcribes and translates audio in many languages.
Whisper ported to C/C++ for fast speech recognition on CPU, Mac and phones without Python.
Music
29Generates complete songs with vocals and instruments from a text prompt or your own lyrics.
Chinese AI company MiniMax has released Music 3, a music model with freely downloadable weights.
Open model that generates long songs with vocals and accompaniment from lyrics.
Open music model that quickly generates full songs with vocals from text and lyrics locally.
AI composer for instrumental music such as film, game and classical scores, with MIDI export.
Command-line tool and Python package for stem separation using UVR models.
Meta's open library including MusicGen, which generates instrumental music from text or a melody.
Creates royalty-free background music matched to the mood of videos and podcasts.
Simple app to generate songs in seconds and publish them to streaming services.
ElevenLabs' music generator that creates songs with vocals from text prompts, also for commercial use.
Platform for AI singing voices, voice conversion, stem separation and mastering for musicians.
Online service to separate vocals, instruments and voices from audio and video files.
Online service for automatic AI mastering of songs, plus distribution and samples.
AI music generator and library of royalty-free tracks for content creators.
Open model from Google Magenta that generates music live and interactively in real time.
App for musicians to separate stems, change tempo, detect chords and practice.
Generates endless or short royalty-free music for videos, streams and apps, also via API.
AI song generator for tracks with vocals from text or reference audio, using its own music model.
AI song generator for tracks with vocals, with generous free usage.
AI generator for royalty-free beats and tracks in many genres, with stem downloads.
Generates royalty-free background music by genre, mood and length that you can customize.
Deezer's open library that quickly splits music into vocals and accompaniment stems.
Stability AI's music and sound generator that creates tracks and effects from text descriptions.
Stability AI's downloadable audio model for short music pieces, loops and sounds.
AI music generator for songs with vocals in many styles, including remix and extend features.
Free desktop app to remove vocals and separate stems using many AI models.
Runs locally
38The easiest way to run AI models like Gemma, Qwen or Llama locally.
Launch a single file and chat locally, with image, speech and music features built in.
A desktop app for local language models: find a model, download it and start chatting offline right away, no terminal required.
Desktop app for local AI: picks the right model for you, chats offline and works with PDFs and files.
Your own ChatGPT as an app – load models with a click, everything stays offline.
Fast C/C++ engine that runs GGUF language models on CPU and GPU.
Your own offline server with Wikipedia, maps and a local AI assistant, and it keeps running without any internet at all.
A local interface that doesn't just run language models but also lets you fine-tune them on your own data
C inference engine that runs huge MoE models without a GPU by streaming experts from the SSD.
Linux kernel module that removes a throttle on RTX 40 GPUs and boosts FP16 tensor throughput.
Tiny C99 engine that runs Kimi K3 on a CPU without a GPU by streaming the model from SSD.
Alibaba framework for running AI models efficiently on phones and other devices.
Compact NVIDIA AI computer with 128 GB memory, clusterable to run large models at your desk.
Web browser whose WebAI preview can run AI models like Gemma locally inside the browser.
Free Mac app that runs AI models for video, images, speech and music offline on Apple Silicon Macs.
Local inference software for running large MoE models like DeepSeek V4.1 Flash on consumer GPUs.
Offline desktop app combining a local AI model with an emergency knowledge library and cited sources.
All-in-one app for local AI chat with your own documents (RAG) and agents.
Microsoft tool to run AI models locally on Windows and macOS with an OpenAI-compatible API.
Desktop app by Nomic to use language models locally and chat with your own documents.
Mozilla project that bundles a model and runtime into a single executable file.
Self-hosted drop-in OpenAI API replacement running text, image and audio models locally.
Engine to run language models natively on phones, in the browser and on various GPUs.
Desktop app for local and cloud models with side-by-side comparison and knowledge stacks.
Browser extension that makes local models (e.g. Ollama) usable in a browser sidebar.
Browser-like app that installs and launches local AI programs with one click.
Local application to query private documents with a language model without internet.
Command-line tool that runs AI models locally inside containers (Podman/Docker).
Fast serving framework to run large language and vision models on your own GPU servers.
Local interface for roleplay and character chats with local or cloud models.
Feature-rich web interface for running language models locally with many backends.
Open models
99Open language models by Zhipu AI (Z.ai), such as GLM-4.5, focused on reasoning and coding.
Small open models from Liquid AI for text, vision and audio that can even run in the browser.
Ant Group's mixture-of-experts model family; Ling-3.2-flash understands text, images and video.
Alibaba's versatile open model family for text, code, vision and more, from tiny to very large.
Large open-weight language model from Mistral that understands text and images.
NVIDIA open model family; Bolt is a compressed vision-language variant for coding and tool use.
Qwen model fine-tuned by Nous Research to avoid repeating Chinese state propaganda on political topics.
Microsoft's efficient 1.58-bit language model that runs with little memory even on weak hardware.
Prism ML vision model family with ternary weights, very small while keeping near-full quality.
Prism ML has released Bonsai 2, a model family built on ternary weights (−1, 0, +1).
Startup PrismML has crushed Qwen3.6-27B down to roughly one bit per weight.
Hugging Face has released Carbon-X, an open model that analyzes DNA sequences in plain language and classifies mutations as benign or …
27B Cloudflare model answering yes/no, choice and scoring questions on text, JSON and images with probabilities.
Open 35B diffusion language model for agent and coding tasks under Apache 2.0 license.
DiffuSpace, a team linked to the Dream models, has released Clover-35B-A3B-Preview as an open diffusion language model.
DeepSeek has released a minor update to its V4.1-Flash model, free to download under an MIT license.
Google has released DiffusionGemma, an experimental open language model that doesn't write text one word at a time.
Italian AI company Domyn began a phased release of Domyn Edge on October 20, 2026.
Open Google embedding model that turns text, code, images, video and audio into vectors for search.
Google DeepMind has released EmbeddingGemma 2, an open embedding model built on Gemma 4.
Baidu's language model family, partly with open weights, such as the ERNIE 5.5 Flash Base model.
Baidu has put its new base model, ERNIE 5.5 Flash Base, on Hugging Face with little fanfare.
Google's compact open language models that run on ordinary computers and phones.
IBM bundles four open Granite time series models into an ensemble that delivers median and quantile forecasts with no retraining.
Open-weight model family derived from DeepSeek; largest variant has 280B parameters and 1M-token context.
"GrootTeam" has released a family of open-weight models.
Hi3D has released its 3D model V1.0, formerly Sparc3D, with code and weights.
Open Tencent model that decides what AI assistants store in long-term memory and how memories get updated.
Chinese company Deep Principle has released Hydrogen-1.0, a specialist model for chemistry, materials science and life sciences built on …
IBM's open time-series models for forecasting without retraining, e.g. TTM and PatchTST-FM.
South Korean tech giant Kakao has released two embedding models on Hugging Face, with 3 billion and 1.3 billion parameters.
Moonshot AI has unveiled Kimi K3.1 Pro, a Mixture-of-Experts model with 2.8 trillion parameters and a 1-million-token context window.
Kolibri is an English-German mixture-of-experts model with 78.1B parameters, of which only 3.46B are active per token.
NVIDIA has introduced Kumo Relational, a pretrained model that makes predictions such as customer churn or fraud directly on relational …
Liquid AI's LFM2.5-24B-A2B is a Mixture-of-Experts model with 24B parameters but only about 2B active per token.
InclusionAI, the AI arm of Alipay operator Ant Group, is now offering its smallest model, Ling 3.0 Tiny, on OpenRouter too – for a tiny …
According to its model card, InclusionAI, Ant Group's lab, has released Ling 3.1 Flash as an open-weight model: 560 billion parameters, …
inclusionAI, Ant Group's AI lab, has released Ling-3.0-flash-Sante, a model specialized in health and medicine, with open weights.
Microsoft has released Mage-VL-Cowork, a small 4-billion-parameter model that controls a computer through the screen and the terminal.
Ternary 20B reasoning model by Deepgrove, under 6 GB and fast on CPUs and small GPUs.
Small open code models from JetBrains, e.g. Mellum2.1 Thinking with 12B parameters for coding agents.
JetBrains has released Mellum2.1 Thinking, a reasoning model with 12 billion parameters, of which only 2.5 billion are active per token.
ModelBest and OpenBMB have released the MiniCPM-OCR series: a 1.3-billion-parameter model that reads scanned and photographed documents, …
Open multimodal model family from MiniMax for coding and agents, with long context.
MiniMax has unveiled M3.1 Flash, a multimodal Mixture-of-Experts model for coding and agents.
Tiny vision model by Cactus Compute that describes scenes, detects objects, reads text and segments on phones.
Open language models from French company Mistral AI, such as Mistral Small and Devstral.
Mistral has released the weights of Mistral Large 4, a model with roughly one trillion parameters that understands text and images.
Taobao's MNN team has open-sourced MNN-Pocket-0.3B, a tiny language model that runs fully offline on your smartphone.
Meta has released Muse Glimmer, an open agent model with 30 billion parameters that runs on a Mac or a PC with a single GPU.
Open language model family whose 33B model targets coding and office agents with 512K context.
The Nanbeige team has released Nanbeige4.3-33B on Hugging Face.
NVIDIA has released Nemotron 3.5 Bolt, a compressed vision-language model that is now available in the Ollama library.
NVIDIA has released Nemotron 3.5 Embed, Rerank and Parse with open weights.
9B model by Bespoke Labs that scores fixed-choice questions with probabilities instead of writing text.
Small 4B NVIDIA language model that loops its layers, giving better results with more compute.
NVIDIA has unveiled Dory, an open 4-billion-parameter language model that runs part of its layers several times over.
NVIDIA's open model family; the 3.5 Super VL variant handles text, images and video with up to 1M tokens of context.
NVIDIA has released Nemotron 3.5 Super VL in a 4-bit variant (NVFP4).
The Allen Institute for AI (Ai2) has released OlmoEarth v1.3, an open model that turns satellite data into compact feature vectors.
Open MoE model family by Ornith AI for coding and agents, runs on a single 80 GB GPU per the maker.
Ornith AI has released Ornith-1.5-122B, a 122-billion-parameter Mixture-of-Experts model built on Qwen3.5.
The team behind Chinese startup SoMark has released OXR, an open-weights model with 1.3 billion parameters that turns PDFs and document …
Japanese AI company Preferred Networks has released its translation model PLaMo 3 Translate as a free download in four sizes, from 0.8 …
Small local model that turns a sentence into a Three.js scene with GLB export for Godot.
A developer has compiled a working 32-bit RISC-V CPU into Qwen 3.5 4B that runs entirely within the model's own computation.
Alibaba's Qwen3.8-27B has just 27 billion parameters, yet in benchmarks it lands close to models 30 to 60 times its size.
A community user shows how a heavily compressed coder variant of Qwen3.8-Flash-Next runs with 262K context on an 8 GB graphics card.
Open KRAFTON language models focused on Korean, reasoning and agents; Raon-1 mini runs on an RTX 5090.
Game company KRAFTON has released Raon-1 mini (Preview), a language model it trained itself, with 28.8B parameters, of which only 4.3B …
US startup Reflection AI has released Beam, a 501-billion-parameter open model under the Apache 2.0 license that it says can match …
The University of Washington's Institute for Protein Design and Nvidia have unveiled RFD4-Proteina, an open generative model for …
US company Ricursive Intelligence has released Ripple-65B Preview, a 65-billion-parameter language model trained from scratch and …
Open 35B MoE agent model for research that can delegate tasks to subagents and verify its own reasoning.
SenseNova has released an open 27-billion-parameter model that doesn't generate video directly.
Open 9B text-and-image model that runs layers with shared weights multiple times.
SenseNova has released the preview model SenseNova-Loop-9B on Hugging Face.
Compact open language models for on-device use, with an adjustable thinking budget.
The SparkLLM team (XHToken) has released Spark-X2.6-4B, a small language model with about 4 billion parameters built to run on your own …
Prism ML has compressed Qwen3.5-9B so hard that the vision-language model now weighs in at just 2.42 GB.
Small open Mistral AI models that forecast time series like sales or sensor data without training.
Mistral AI introduces Timestral, a family of small foundation models for time-series forecasting.
Baidu has released Unlimited OCR 2, a 3.34-billion-parameter OCR model under the MIT license.
Open video model (9.5B parameters) for ComfyUI that creates clips from text or animates images.
Xiaomi has released Xiaomi-OCR-0, an open-source document model with just 0.8 billion parameters.
China Telecom AI has released Xing4.0-29B-A4B, an open language model built to work through multi-step tasks and call tools on its own.
Open vision-language-action model by Physical Intelligence for controlling robots.
Open language models from the Technology Innovation Institute (Abu Dhabi), incl. Falcon 3 and Falcon-H1.
OpenAI's open-weight reasoning models in two sizes (20B and 120B) for local use.
IBM's open language models for business use, code and document processing.
Meta's open language model family, available in many sizes for download and local use.
Open language models by MiniMax with very long context and strengths in agents and code.
Fully open language models from Ai2, including training data and training code.
Open language model from ByteDance Seed with long context and reasoning abilities.
Tiny open language models from Hugging Face that even run in the browser or on phones.
For developers
45Very fast diffusion language model from Inception, available via API and chat.
Google's method for invisibly watermarking AI text, images and audio; the text version is open source.
Open standard for giving AI agents capabilities as folders of instructions and scripts.
AWS service offering AI models from multiple providers through a single interface.
Cohere API model that re-sorts search results by relevance before a language model uses them.
Anthropic research prototype that uses zero-knowledge proofs to show which AI model produced an answer.
Tool that removes refusals and other behaviors from local language models without training.
Platform for sharing and using AI models, datasets and demos (Spaces).
NVIDIA open-source framework for training robots with reinforcement learning in simulation.
Open-source proof assistant and programming language for machine-checking mathematical proofs.
Open-source NVIDIA toolkit for building, connecting and monitoring enterprise AI agents.
Platform of open world models and tools for robotics and autonomous vehicle AI.
Open-source sandbox runtime where AI agents access files, network and credentials only per defined policies.
Lightweight framework for multi-agent workflows from OpenAI.
Open-source Anthropic tool that restricts file and network access of processes without containers.
Open-source NVIDIA scanner that checks AI agent skills for malicious code and prompt injection before installation.
Tiny world model that generates and plays simple games on a Raspberry Pi Pico 2 W.
Google's open-source PyTorch backend that runs PyTorch models natively on TPUs.
Anthropic's API for the Claude models.
Web interface and API for trying and using Gemini models.
Cloud API offering very fast inference for open language models.
Framework for connecting your own data to language models (RAG, agents).
Open standard for connecting AI assistants to tools and data sources.
Unified API for hundreds of language models from many providers.
Managed cloud vector database for AI search.
Service for running AI models in the cloud via API.
Cloud platform to run and fine-tune open models via API.