AI RoadAI Open Source

AI Open Source

Hand-picked AI open-source projects on GitHub worth following: LLMs, agents, RAG, local deployment, image and speech… 1056 projects so far.

Looking for projects you can deploy for free on Cloudflare, Vercel, Netlify and more? See Free Deploy Picks

All 1056🧠 Large Language Models 26🤖 AI Agents 166🧩 App Development Platforms 23📚 RAG & Knowledge Bases 46💬 Chat UIs & Clients 42🖥️ Local Inference & Deployment 87👨‍💻 AI Coding 99🛠️ Frameworks & SDKs 48🔌 MCP & Tool Ecosystem 35🎨 Image Generation 25🎬 Video Generation 17🎙️ Speech & Audio 43👁️ Computer Vision 28📄 Document Parsing & Scraping 25🗄️ Vector Databases 25🏋️ Training, Fine-Tuning & MLOps 76🧪 Evaluation, Observability & Safety 82🚪 API Gateways & Proxies 12🔎 Search & Deep Research 15⚙️ Deep Learning Libraries 57🏷️ Data & Annotation 32🔬 Industry & Science AI 35⚡ Productivity Tools 12
1047 projectsSort: RecommendedStarsMost visitedNewest

Switchyard lets LLM applications route traffic across models and providers while preserving native OpenAI and Anthropic API compatibility - enabling flexible model selection, benchmarking, and cost/performance optimizati

Run Claude Code 100% on-device with local AI on Apple Silicon. MLX-native Anthropic-API server. 6 fighters incl. Muse-Glimmer 30B (now multimodal — reads images, abliterated), Gemma 4 31B, Qwen 3.5 122B (65 tok/s), DeepS

🖥️ Local Inference & DeploymentPythonLlamaLocal deployment★ 3.4k

A single Gradio + React WebUI with extensions for ACE-Step, OmniVoice, Kimi Audio, Piper TTS, GPT-SoVITS, CosyVoice, XTTSv2, DIA, Kokoro, OpenVoice, ParlerTTS, Stable Audio, MMS, StyleTTS2, MAGNet, AudioGen, MusicGen, To

🎙️ Speech & AudioTypeScript★ 3.3k

The Swiss Army Knife of Offline AI. Chat, see, speak, and generate images on your phone or Mac — GGUF LLMs, vision, Whisper speech-to-text, Stable Diffusion, tool calling, and local-network servers. Runs on your CPU, GPU

🎨 Image GenerationTypeScriptMCP★ 3.3k

Open-source, self-hosted AI penetration testing framework: maps your attack surface into a graph, autonomously exploits it from a Kali sandbox with human approval gates, and opens PRs that fix what it finds. MCP both way