An open-source agent workspace for knowledge work and software development that allows users to automate workflows and build AI tools using a choice of models and providers.

A comprehensive digital human framework that integrates LLMs, ASR, and TTS to power interactive virtual avatars for apps, websites, and embedded devices.

A local-first coding agent that automates project planning, implementation, and verification from the command line using Git worktrees and local project memory.

A reference implementation of the Jacobian lens for decoding internal language model activations into vocabulary tokens to analyze internal representations.

transcribe.cpp is an MIT‑licensed C/C++ library that runs 60+ speech‑to‑text models (Whisper, Parakeet, Granite, Voxtral, etc.) via the GGUF format on Metal, Vulkan, CUDA, HIP/ROCm, or a tinyBLAS‑accelerated CPU path. It provides a CLI, a quantization tool, conversion scripts from NVIDIA NeMo, and official Python, TypeScript, Rust, and Swift bindings. Models are hosted on Hugging Face and are numerically verified with WER tests.

A Deep Reinforcement Learning library providing single-file, transparent implementations of popular algorithms to simplify research, debugging, and prototyping.

Serenity.skill is an AI‑agent plug‑in that encodes a structured “Serenity” research method for turning market hot‑topics into a short, evidence‑backed list of stocks or funds. Install by copying the folder into a Codex or Claude Code skill directory, then call it with natural‑language prompts (e.g., `$serenity‑skill 研究现在 A 股 AI 半导体产业链`). The skill generates a step‑by‑step analysis—splitting a theme into supply‑chain layers, identifying bottlenecks, pulling public filings, and ranking candidates—while explicitly stating evidence gaps and risks. It does not trade or provide live data; it merely assists the user’s research workflow. MIT licensed.

An open-source data quality tool that uses 'Expectations' as unit tests for data to validate quality and automatically generate documentation.

ClawGod is a GPL‑licensed wrapper that extracts the official Claude Code binary, applies a set of patches to unlock hidden commands, remove safety restrictions, and add custom behaviour, then runs the patched CLI under Bun. It auto‑updates, offers per‑feature toggles, and works on macOS, Linux, and Windows.

A collaborative speedrun project focused on finding the fastest possible algorithm to train a language model to a specific performance target on 8x H100 GPUs.

A GPU-accelerated, lightweight learning library for robotics research that enables rapid prototyping of robot learning algorithms.

A plugin that connects AI agents like Codex and Claude Code to ChatCut, enabling users to edit video projects, generate assets, and manage timelines using natural language.

A Dockerized FastAPI wrapper for the Kokoro-82M TTS model that provides an OpenAI-compatible API for high-quality, multi-language speech generation.

A lightweight terminal coding assistant that implements a Claude Code-like agentic workflow for local development, focusing on simplicity and extensibility.

A self-hosted API relay service for Claude that enables multi-account management, cost-sharing, and private access to Anthropic's models.

A cloud-native, Kubernetes-native control plane for deploying and managing scalable and cost-effective LLM inference infrastructure.

A multi-physics simulation package for modeling and simulating rigid and deformable bodies, granular dynamics, and fluid-solid interactions, specifically supporting robotics and embodied AI.

A full-stack engineering platform for deploying embodied intelligence applications, providing a standardized loop from VLA model training to real-robot deployment.

A local AI agent desktop application for macOS, Windows, and Linux that packages the Pi agent's capabilities into a standalone app with session management and project browsing.

A Codex-native SEO audit suite that orchestrates specialist AI agents and deterministic runners to perform technical, content, and AI-search (GEO) analysis.

A suite of AI agent skills for calculating and analyzing Vedic astrology charts directly from birth details, featuring specialized modules for career, love, and birth time rectification.

A collection of developer utilities and demos for the WebMCP API, enabling web pages to expose structured tools for AI agents to interact with programmatically.

A Go implementation of the Model Context Protocol (MCP) that enables LLM applications to integrate with external data sources and tools using a standardized interface.

The official Go SDK for the Model Context Protocol (MCP), providing the tools to build standardized clients and servers for AI model interaction.

An open source platform for running managed agents that handles execution and state, allowing developers to to choose their own models, tools, and machines.

The GitHub Copilot app is a desktop client that lets you run Copilot‑powered AI agents in isolated local or cloud workspaces, manage multiple parallel sessions, and oversee the entire coding workflow—from picking a task to merging a PR—through visual canvases and automation tools.

An open-source AIOps and alert management platform that centralizes, deduplicates, and uses AI to correlate and summarize alerts from multiple observability tools.

A Python library that automatically learns web scraping rules from sample data, eliminating the need for manual CSS selectors or XPath expressions.

An AI-powered translation tool that creates bilingual versions of EPUB, PDF, TXT, and other documents using LLMs and machine translation engines.

A self-hosted AI transcription and note-taking platform that transforms audio recordings into searchable, intelligent notes with privacy-first infrastructure.

A high-performance RLHF framework using Ray and vLLM to scale reinforcement learning for LLMs and VLMs, featuring a unified agent-based design for single and multi-turn training.

A collection of Model Context Protocol (MCP) servers that enable AI agents to manage Ubiquiti UniFi Network, Protect, and Access controllers via natural language.

A local proxy that bridges the Claude Agent SDK to the standard Anthropic and OpenAI APIs, allowing third-party AI agents to use Claude Max subscription access.

A curated collection of papers, frameworks, and tools for on-policy distillation, a post-training technique that trains student LLMs on their own generated samples to bridge the train-inference distribution gap.

An AI-powered research and writing tool that generates publication-ready scientific papers, grant proposals, and reports backed by real-time literature search and verified citations.

Maestro is a cross‑platform desktop app that lets power users run, automate, and monitor many AI coding agents (Claude Code, OpenAI Codex, etc.) in parallel, with Git worktree support, playbook automation, a keyboard‑first UI, remote web control, and analytics—all under an AGPL‑3.0 license.

An MCP server that enables AI agents to perform professional database administration, including index tuning, health checks, and query optimization for PostgreSQL.

A plugin and MCP server that allows AI agents to read and edit Figma documents in real-time while bypassing official API rate limits.

A web extraction tool that converts websites into clean Markdown, JSON, and LLM-ready context for AI agents and RAG pipelines.

A Codex-only workflow for capability-routed software delivery that automates planning, implementation, and verification using risk-based agent routing.

An infrastructure layer that provides a single, shared persistent memory database with vector search for all your AI tools, eliminating fragmented data across different AI clients.

An AI‑agent skill that audits a repository, proves that dead or redundant code can be safely removed, and applies the change with full evidence and rollback support. It works in read‑only “Survey” mode or authorized “Change” mode, offers optional interactive HTML reports, and is packaged for Codex‑compatible agents under an MIT license.

A comprehensive set of AI agent skills and modules that transforms LLMs into a professional marketing engine for audits, copy generation, and asset production.

A collection of SQL and AI-based models for detecting financial fraud, credit risk, and regulatory compliance issues in lending and corporate finance.

A ComfyUI plugin that decomposes single anime illustrations into layered 2.5D models with depth ordering, specifically designed for Live2D workflows.

Nav2 is the official ROS 2 navigation stack, offering localisation, path planning, obstacle avoidance, and docking for mobile robots. It’s modular, supports multiple ROS 2 releases, and is backed by a strong community and industry sponsors.

A remote Model Context Protocol (MCP) server that transforms any public GitHub repository into a documentation hub for AI assistants to reduce hallucinations.

An enterprise-grade open-source RPA desktop application that enables low-code automation of desktop software and web pages, featuring deep integration with AI agents.