Explore the directory

A kernel library and generator for LLM inference that provides unified, high-performance attention, GEMM, and mixture-of-experts kernels across GPU backends.

• Inference

An open-source Python library for offline machine translation, usable as a library, a command-line tool, or a GUI, with automatic pivoting through an intermediate language when no direct model is installed.

• Libraries

An open-source, self-hosted cloud agent cluster that runs a collaborative team of AI agents able to call each other and connect to social platforms and Web3.

• Agent frameworks

A Kubernetes operator that creates and autoscales self-hosted GitHub Actions runners as scale sets, so ephemeral, container-based runners can scale up or down with running workflows.

• CI/CD
• Containers

Google's open-source orchestration layer for running large language models on edge devices, built for cross-platform, hardware-accelerated on-device inference.

• Inference

A Grok-powered terminal AI agent that reads and edits its own codebase, with a Model Context Protocol integration for custom tools and sandboxed file and shell access.

• Coding agents

A curated list of client applications that implement the Model Context Protocol, covering desktop apps, CLI tools, editor integrations, and chatbots.

• MCP servers
• Curated lists

A collection of Triton kernels for LLM training, from LinkedIn, that reduces GPU memory use and increases training throughput with a one-line change.

• Models

An IntelliJ IDEA plugin that provides a visual chat and agent interface for Claude Code, OpenAI Codex, and several other terminal-based AI coding CLIs.

• Coding agents
• Editors

A KV-cache-centric, disaggregated serving architecture for large language models, developed as the platform behind Moonshot AI's Kimi service.

• Inference

An unofficial MCP bridge that connects Claude Code to a locally running TradingView Desktop app over Chrome DevTools Protocol for chart analysis and Pine Script work.

• MCP servers

A Python package and command-line tool that crawls, downloads, and extracts the main text, metadata, and comments from web pages into CSV, JSON, HTML, Markdown, or XML.

• Browser automation

A command-line tool that detects a machine's GPU, CPU, and RAM, then ranks which local language models actually run well on it using live benchmark data.

• Inference

A research toolkit that compresses LLM prompts and key-value caches by removing non-essential tokens, reporting up to 20x compression with little quality loss.

• Inference
• Libraries

A Go library for building a customizable HTTP and HTTPS proxy server, supporting request and response manipulation, MITM certificate caching, and hijacked connections.

• Libraries
• Networking

A Codex skill that uses an already-paid ChatGPT web subscription as the planning and review brain for a Codex coding session, through a read-only MCP bridge.

• Coding agents
• Agent skills

A Chinese-language textbook on large models and agents paired with six progressive programming exercises, from a hand-written Transformer through a coding agent.

• Tutorials
• Books

A local-first agent workbench that unifies multiple agent engines, such as Codex and OpenCode, plus plugins, skills, and multi-agent projects, in one workspace.

• Coding agents

An AI coding agent skill that scaffolds a full Next.js editor for producing App Store and Google Play marketing screenshots at every required device size.

• Agent skills
• Design tools

A custom Swift and Metal runtime that runs the 26-billion-parameter Gemma 4 26B-A4B model in about 2 GB of RAM on any Apple Silicon Mac, including 8 GB models.

• Inference

A conversational data analysis system that turns natural-language questions into SQL using a large language model and retrieval-augmented generation.

• Databases
• Data engineering

An open-source TypeScript framework for building agents that share a purpose-built UI, so the same action can be called by the agent as a tool or by application code.

• Agent frameworks

A self-hosted deployment system that installs and wires together a private AI server, combining local model inference, a chat web UI, automation, and search.

• Inference
• Hosting

A local browser interface for the pi coding agent that reuses its session and configuration files to browse, resume, and run agent conversations.

• Coding agents