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.
310 projects
sparkjsdev
An advanced 3D Gaussian Splatting renderer for THREE.js with cross-device WebGL2 support, multi-format loading, dynamic editing, and streaming LoD.
pydantic
Type-safe Python GenAI agent framework by the Pydantic team with model-agnostic support, structured outputs, MCP/A2A integration, and production observability.
speaches-ai
OpenAI API-compatible self-hosted speech server for streaming transcription, translation, and speech generation — Ollama for TTS/STT models.
openvinotoolkit
Intel's open-source AI inference optimization toolkit supporting PyTorch, TensorFlow, and ONNX across CPU, GPU, and NPU hardware with INT8/FP16 quantization.
roboflow
Real-time transformer architecture for object detection and instance segmentation with DINOv2 backbone, achieving state-of-the-art COCO performance from Nano to 2XL model scales.
dimensionalOS
Agentic operating system for generalist robotics enabling natural language control of humanoids, quadrupeds, and drones with SLAM, spatial memory, and MCP support.
langchain-ai
Batteries-included AI agent harness built on LangChain/LangGraph with planning, filesystem operations, sub-agent spawning, and context management for production-ready agent deployments.
topoteretes
Open-source knowledge engine for persistent AI agent memory combining vector search, graph databases, and cognitive science for unified knowledge infrastructure.
FireRedTeam
Industrial-grade ASR system achieving SOTA 3.05% CER on Mandarin benchmarks with two model variants (8.3B LLM and 1.1B AED), supporting Chinese dialects, English, and singing lyrics recognition.
Gen-Verse
Open-source multimodal diffusion language model family unifying text reasoning, visual understanding, and image generation through block diffusion, mixed CoT, and UniGRPO reinforcement learning.
OpenMOSS
Comprehensive open-source speech and sound generation family with 5 production-ready models (1.7B-8B) covering TTS, dialogue, voice design, real-time streaming, and sound effects across 20 languages.
QwenLM
Open-source multilingual TTS from Alibaba's Qwen team with 97ms streaming, 3-second voice cloning, and natural language voice design across 10 languages.