Skills, agents, hooks. Cross-harness. With a side of Pi.

Hundreds of models & providers. One command to find what runs on your hardware.

A simple TUI for serving local LLM models. Pick a model, pick a backend, serve it

MiniJinja is a powerful but minimal dependency template engine for Rust compatible with Jinja/Jinja2

Golang inference engine and deep learning primitives

A standalone terminal agent in a single C file. Model agnostic (13 providers + local hako), skill-driven, tool permission gated. No node, no python.

C inference engine; own loader, kernels, and tokenizer. Runs hako-sho (3B) and hako-koi (7B) on CPU. No llama.cpp, no ollama.

A harness optimized to smaller LLMs

CLI tool for interacting with the web

AI coding agent optimized for small LLMs. 87% benchmark with 4B-active model.

Community recipes for serving LLMs on RTX 3090/4090/5090 CUDA gpus. Multi-engine (vLLM, llama.cpp, ik_llama) and model-agnostic. Currently shipping Qwen3.6-27B Qwen3.6 35B Gemma 4 26B Gemma 4 31B c…

A fast Node.js package manager

The Pi desktop app you want to use.

Native desktop shell for Pi Coding Agent (Tauri + Lit), extension-first and multi-session aware

Persistent JavaScript REPL for context manipulation, MCP integration, and focused sub-agent orchestration. A pi extension.

🕶️ Parse GitHub Actions inputs from both `workflow_dispatch` and `repository_dispatch` events

Manage AppImages with ease 📦

A workflow booster for solving CI failures in moonrepo repositories

Open-source React component library. Polished, composable components for React and Tailwind CSS. Distributed via a shadcn-compatible registry. Never paywalled.

AI agent toolkit: unified LLM API, agent loop, TUI, coding agent CLI

Multi-agent communication extension for pi coding agent

Pi extension for async subagent delegation with truncation, artifacts, and session sharing

code intelligence on your CLI for AI agents. tell your agent to run `cm --help` and let it rip

Experimental Linux microvm setup with a TypeScript Control Plane as Agent Sandbox

Analysis

VS Code-style command palette for selecting and applying skills in pi

mini cli search engine for your docs, knowledge bases, meeting notes, whatever. Tracking current sota approaches while being all local

Lightweight WASM-based Linux VM for AI agents - run shell commands in an isolated Alpine Linux environment

Retain is a native macOS application that aggregates AI conversations from multiple platforms into a unified, searchable knowledge base with intelligent learning extraction.

A self-learning system for Claude Code that captures corrections, positive feedback, and preferences — then syncs them to CLAUDE.md and AGENTS.md.