Explore the directory

A full-stack codebase for training and evaluating draft models used in speculative decoding, covering data preparation, training and benchmark evaluation.

• Models

A trilingual (Traditional Chinese, English, Simplified Chinese) learning roadmap for agentic AI, with more than 240 curated resources and hands-on exercises from LLM basics to multi-agent systems.

• Tutorials

A self-hostable Git server for the command line, with an SSH-based text UI for browsing repositories, access control by public key, and Git LFS support.

• Git

An open-source, document-centric desktop app for working with AI agents, built on the Claude Agent SDK and the Pi SDK, that connects to APIs, MCP servers and existing Claude Code skills.

• Coding agents

An enterprise-grade, production-ready framework for orchestrating multiple AI agents, with prebuilt sequential, concurrent and hierarchical multi-agent architectures.

• Agent frameworks

A self-hosted, multi-tenant knowledge agent platform combining RAG, a knowledge graph, multi-agent orchestration, MCP/Skills, a sandbox and permission management in one workspace.

• Agent frameworks
• RAG

A C++ library for parsing and generating JSON and MessagePack on Arduino boards and other embedded and desktop C++ projects.

• Libraries

A collection of agent skills covering UI, typography, color, accessibility, layout and product writing, to help a coding agent build and review a good interface.

• Agent skills
• Design tools

A skill store aggregating official and community AI agent skills across document processing, content creation, programming, machine learning and automated workflows, auto-updated daily.

• Agent skills

An agent distro that lets a single coding-agent session supervise a crew of other agents, each working in its own git worktree, and report back finished pull requests, approved merges, or investigation reports.

• Coding agents

A family of open-weight language models from Z.ai focused on coding and long-horizon agentic tasks, with GLM-5.3 adding stronger coding and emergent cyber capability.

• Models

An LLM-driven platform that extracts structured JSON from unstructured documents such as PDFs, scans and images, deployable as an API or an ETL pipeline.

• RAG
• Data engineering

A Claude Code plugin marketplace from Trail of Bits providing security-research skills for smart contract auditing, code review, and other security analysis workflows.

• Agent skills
• Security

A Claude Code / Codex skill that turns an article, screenshot or video into Xiaohongshu image carousels and WeChat cover-image pairs, rendered from single-file HTML to PNG.

• Agent skills
• Generative media

A community-driven registry service, like an app store for Model Context Protocol servers, giving MCP clients a searchable list of available MCP servers.

• MCP servers

An MCP server and Chrome extension that streams console output, network activity, screenshots and Lighthouse audits from a developer's real, logged-in browser session to any MCP-compatible AI client.

• MCP servers
• Debugging

A deterministic engineering layer that configures existing AI coding agents with persistent memory, a structured development workflow, and evidence-based, bounded review.

• Coding agents

A lightweight, open-source MCP server that bridges AI agents and the Windows operating system, for file navigation, application control, UI interaction and QA testing.

• MCP servers
• Testing

An open-source, all-in-one WebUI for AI image and video generation, captioning and processing, built on Stable Diffusion and supporting dozens of models.

• Generative media

A free, ten-week, hands-on course on building production-ready LLM applications, covering RAG, vector search, embeddings, agents, function calling, evaluation and monitoring.

• Tutorials

An AI-powered dark web OSINT tool that uses an LLM to refine search queries, filter results from dark web search engines, and produce an investigation summary.

• Security

Memory infrastructure for building stateful AI agents that understand people, other agents, groups and projects as they change over time, usable managed, local or self-hosted.

• Agent memory

A small family of yes/no, multiple-choice and rating decision models, built on Qwen3.5 and Qwen3.8, that can be trained, fine-tuned and served from your own hardware.

• Models

A toolkit for merging pre-trained large language models using an out-of-core approach that can run entirely on CPU or with as little as 8 GB of VRAM.

• Models