Open Source
Explore the latest AI open-source projects from GitHub and HuggingFace.
Explore the latest AI open-source projects from GitHub and HuggingFace.
Explore the latest AI open-source projects from GitHub and HuggingFace. Discover projects across categories like LLM, Vision, Audio, and more.
886 projects
ruvnet
An MIT-licensed, Rust-based platform that turns commodity WiFi Channel State Information from cheap ESP32 boards into contactless sensing — presence, vital signs, pose, and activity — with Home Assistant and Matter integration.
dottxt-ai
A mature Apache-2.0 Python library for structured LLM generation that constrains output at the token level to match Pydantic models, JSON, regex, choices, or grammars, working across local models, servers, and hosted APIs.
microsoft
A Microsoft open-source optimizer that trains reusable natural-language skills for frozen LLM agents through trajectory-driven, validation-gated edits, producing compact best_skill.md artifacts with no added inference cost.
RunanywhereAI
A source-available toolkit for running LLMs, vision, speech-to-text, and text-to-speech models fully on-device across iOS, Android, Flutter, React Native, and Web, with Qualcomm Hexagon NPU acceleration.
1jehuang
A next-generation, Rust-built coding agent harness focused on multi-session workflows, low resource use, and semantic agent memory, with cross-platform installs and provider-agnostic model support.
Tencent-Hunyuan
Tencent Hunyuan's open text-to-3D human-motion model: a Diffusion Transformer with Flow Matching that generates skeleton-based character animations from text prompts, scaled to the billion-parameter level with 1.0B and 0.46B variants.
DrewThomasson
An open-source, Apache-2.0 tool that converts e-books into chaptered audiobooks using multiple TTS engines, with optional voice cloning, OCR input, and support for 1158 languages on CPU or GPU.
NVIDIA
NVIDIA's open-source, Apache-2.0 speech AI framework behind the Parakeet and Canary ASR models, delivering benchmark-leading speech recognition with controllable-latency streaming, multilingual support, and open weights.
rerun-io
An open-source, Rust-built data layer for physical AI that logs, visualizes, queries, and streams multimodal data (images, point clouds, video, time series) into training, with Python, Rust, and C++ SDKs.
HeartMuLa
An open-source music generation stack that turns lyrics and style tags into full multilingual songs, shipping a 3B model plus its own codec, Whisper-based transcriptor, and audio-text alignment model. Apache-2.0.
Robbyant
A feed-forward 3D foundation model that reconstructs scenes from streaming video in real time, using a Geometric Context Transformer to stay stable across 10,000+ frames. Apache-2.0, built on DINOv2 and VGGT.
Canner
Open-source generative BI engine that turns natural-language questions into governed text-to-SQL, charts, and dashboards through a reviewable context layer, built for AI agents across 20+ data sources.