Run your own AI cluster at home with everyday devices 📱💻 🖥️⌚

Self-hosted AI audio transcription

Audio Editor

Open-Source Frontier Voice AI

🔥🔥 Open-sourced unified customization model

[ICCV 2025] Enhancing spatial understanding in text-to-Image diffusion models

[3DV 2026 Oral] VoxHammer: Training-Free Precise and Coherent 3D Editing in Native 3D Space

​​Unlimited-length talking video generation​​ that supports image-to-video and video-to-video generation

[NeurIPS 2025] Let Them Talk: Audio-Driven Multi-Person Conversational Video Generation

Hierarchical Reasoning Model Official Release

Autonomous, agentic, creative story writing system that incorporates stored embeddings and Knowledge Graphs.

RTX 5090 & RTX 5060 Docker container with PyTorch + TensorFlow. First fully-tested Blackwell GPU support for ML/AI. CUDA 12.8, Python 3.11, Ubuntu 24.04. Works with RTX 50-series (5090/5080/5070/50…

Implement a ChatGPT-like LLM in PyTorch from scratch, step by step

Ollama client for iOS, Android, macOS, and Windows that simplifies experimenting with LLMs.

An AI-powered task-management system you can drop into Cursor, Lovable, Windsurf, Roo, and others.

🙌 OpenHands: AI-Driven Development

A natural language interface for computers

🔥 Official Firecrawl MCP Server - Adds powerful web scraping and search to Cursor, Claude and any other LLM clients.

Context7 MCP Server -- Up-to-date code documentation for LLMs and AI code editors

unsloth-5090-multiple

Fine-tuning & Reinforcement Learning for LLMs. 🦥 Train OpenAI gpt-oss, DeepSeek-R1, Qwen3, Gemma 3, TTS 2x faster with 70% less VRAM.

Fast and memory-efficient exact attention

union filesystem using fuse

This package allows macOS Finder to display thumbnails, static QuickLook previews, cover art and metadata for most types of video files.

Updated November 12th - The original free MacOS plugin to play .webm movies in Quick Look Preview.

A minimalist web native image viewer.

ComfyUI-SimpleLogger is an extension designed to save ComfyUI history in html file.

HunyuanCustom: A Multimodal-Driven Architecture for Customized Video Generation

[CVPR 2025] Official PyTorch implementation of "EdgeTAM: On-Device Track Anything Model"