Explore the directory

A Vim plugin for local LLM-assisted code completion, using fill-in-the-middle suggestions and instruction-based editing against a llama.cpp server, with configurable context scope and named server profiles.

• Editors

The official Python SDK for the Agent2Agent (A2A) protocol, for building and serving agentic applications that communicate over a standardized multi-agent interoperability protocol.

• Agent frameworks

An open-source desktop AI agent workspace that combines everyday task work, prompt-to-editable visual design, and bringing an external coding agent such as Codex or Claude Code into a real project, all in one interface.

• Agent frameworks
• Design tools

A Python library that provides one unified interface for calling any LLM provider, such as OpenAI, Anthropic, Azure, Mistral or Ollama, so switching providers is a one-line change rather than a code rewrite.

• LLM gateways

A free desktop app that assembles teams of AI coding agents across tools like Claude Code, Codex, OpenCode and Cursor, coordinating their work on a kanban board with agent-to-agent review, token budgets, and cross-team communication.

• Coding agents
• Desktop

A set of portable Agent Skills that teach AI coding assistants such as Claude, Copilot, Codex and Cursor expert-level WordPress development: Gutenberg blocks, block themes, plugin architecture, the REST and Abilities APIs, WP-CLI and performance profiling.

• Agent skills

An open-source desktop app that wraps Claude Code and other AI models in a one-click-install GUI with VM-level sandbox isolation, a document-generation Skills system, MCP integration and remote control through Feishu or Slack.

• Coding agents
• Desktop

An open-source LLM inference engine for agentic workloads that separates a C++ control plane, which manages request lifecycles and KV cache safety at compile time, from a Python execution plane for fast iteration.

• Inference

A set of helper libraries for NestJS that generate fast runtime validators, a typed client SDK with a mockup backend simulator, automatic Swagger docs and end-to-end test functions from pure TypeScript types, plus an LLM function-calling library.

• APIs
• Backend

A native iOS and Android app for chatting with a self-hosted Open WebUI server, direct OpenAI-compatible and Ollama endpoints, or on-device Apple Intelligence models, built to fix the mobile gaps in Open WebUI's own web client.

• Inference

Node.js bindings for llama.cpp that run language models locally on Metal, CUDA or Vulkan hardware, with pre-built binaries, JSON-schema-constrained output, function calling and embedding support.

• Inference
• Libraries

A Python tool and Claude Code skill that lets an LLM actually watch a video: it extracts scene-aware, deduplicated keyframes and a transcript from a URL or local file, running entirely on the local machine.

• Agent skills

A local-first AI agent desktop client that can read and write files, run shell commands and manage scheduled tasks on your machine, extendable through MCP servers and Skills packages, with an optional remote gateway for browser access.

• Agent frameworks
• Desktop

Freezed is a Dart code generator that produces immutable data classes, tagged unions and cloning support, so a developer only has to write the definition of a model instead of its boilerplate.

• Libraries

A complete Chinese-language tutorial for OpenClaw, a command-line AI assistant framework, covering installation, configuring skills for everyday workflows, and building a custom version from its source.

• Tutorials

A monospaced coding typeface based on Nanum Barun Gothic, tuned for legibility and to keep similar-looking Latin, Hangul and punctuation characters distinguishable while writing code.

• UI kits

A desktop app or headless server that browses, searches and analyzes conversation history from more than thirty AI coding assistants, including Claude Code, Codex CLI, Cursor and Gemini CLI, entirely offline.

• Coding agents

A set of agent skills that let an orchestrating coding agent discover installed coding-agent CLIs, organize them into named lanes, and delegate a task to the right implementer, while the developer still reviews the diff and lands the commit.

• Coding agents
• Agent skills

An open-source platform that runs every MCP server in an isolated container, enforces identity and access policy per request, and gives platform teams a Kubernetes operator and audit logs for running MCP in production.

• MCP servers
• Containers

An open-source React component library, built on Ant Design, for building AI-generated-content (AIGC) web applications such as chat and generative interfaces.

• UI kits

A terminal coding agent CLI built on the pi-mono agent SDK that can delegate different task types to different models through a multi-model orchestrator, backed by kimchi's own LLM infrastructure.

• Coding agents

A technique and library for distilling any sentence-transformer model into a small, fast static embedding model, cutting size by up to 50 times and inference time by up to 500 times for a small drop in accuracy.

• RAG
• Models

A self-managing memory system for coding agents that captures durable project knowledge during context compaction, consolidates and verifies it against the codebase overnight, and recalls the right memories automatically each turn.

• Agent memory

An open-source testing and evaluation platform for MCP server developers that emulates 16 client configurations and over 170 models, runs evals against test cases, and gates regressions in CI/CD.

• MCP servers
• Testing