Explore the directory

A deterministic, on-device command-line translator for macOS Tahoe that also runs as an HTTP server compatible with the DeepL, LibreTranslate and Google Translate v2 APIs, with no cloud calls or LLM.

• CLI tools
• APIs

The official distribution repository for Supabase's skills and MCP adapters for AI coding assistants, bundling general Supabase and Postgres-performance guidance for Claude Code, Cursor, Codex and others.

• MCP servers
• Agent skills

A self-contained terminal AI coding assistant that speaks Anthropic's Claude API and any OpenAI-compatible endpoint, running an autonomous tool-using loop behind a permission system.

• Coding agents

A persistent JavaScript REPL extension for the pi agent that spawns async subagents in their own tmux sessions and git worktrees, with a full MCP client built in.

• Coding agents

A fine-tuning SDK that captures an AI agent's real traffic, trains a small open model to match it, and gradually shifts the agent from a costly frontier model to the model you now own.

• Models

An autonomous agent that runs the operational side of maintaining open-source repositories, triaging issues and reviewing pull requests across iterations while leaving merges and design decisions to the maintainer.

• Coding agents
• Git

A dependency-free, privacy-first JavaScript library that turns a phone's blowing, tilting, shaking and long-press gestures inside a progressive web app into structured events for any LLM chat backend.

• Libraries

A Chinese-language open-source book for AI engineers and architects on building production multi-agent systems, covering tool calling, coordination, context engineering, safety, evaluation and delivery across ten chapters.

• Books

An OpenAI-compatible gateway for attested confidential inference that publishes hardware attestation for itself, verifies its upstream model providers before forwarding a prompt, and signs a receipt for every response.

• LLM gateways
• Security

A transport layer that lets a robot's cameras, joint state and control actions travel over WebRTC to a teleoperator or policy running on another machine, fusing video and state into one synchronized callback per control tick.

• Networking

Fine-tunes open-source language models using Tinker from inside the Pi agent, preparing training data, running small jobs and comparing the trained checkpoint against the original model.

• Models

A pure-Rust command-line tool that splits PDFs into chunks for retrieval-augmented generation by analyzing layout, font size and boldness rather than a fixed token or character count.

• RAG

A demo multi-agent travel-planning assistant built with LangGraph and MCP, showing a supervisor agent, input guardrails and human-in-the-loop approval around a FastAPI web frontend.

• Agent frameworks
• Tutorials

Captures a team's real engineering conventions, rejected approaches and edge cases as versioned Markdown decision files in a repository, so a coding agent consults them before writing code.

• Agent memory

A local-first memory layer for AI assistants that runs an MCP server over a local SQLite file, so tools such as Claude Code, Cursor and GitHub Copilot can share durable memory about a user, their work and past decisions.

• MCP servers
• Agent memory

A scaffolding CLI and starter kit for building multi-agent applications with LangGraph, offering seven agent patterns, six LLM providers and an HTTP server with streaming out of the box.

• Agent frameworks

A local-first, multi-agent software-engineering runtime with its own scheduler, sandbox, permission system and REPL, running agentic coding workflows through Ollama, llama.cpp, vLLM or LM Studio.

• Coding agents

A local runtime for AI agent tool chains that wraps MCP servers, REST APIs and CLI commands behind one configuration, injects credentials from an encrypted vault, and logs every call.

• Agent frameworks
• Auth

A minimal, protocol-only Python framework that adds tool use, conversation memory and any LLM provider to an existing application, kept to roughly 2,000 auditable lines with one runtime dependency.

• Agent frameworks

Splits one large language model across a Mac and a separate Windows or Linux PC over an encrypted, framework-neutral network protocol, producing output identical to running the whole model on one machine.

• Inference

A TypeScript SDK for the Day AI CRM platform that exposes its MCP tools as a plain API for searching contacts, creating opportunities and pulling meeting transcripts.

• APIs

A Swift package and compiler for authoring AI pipelines as a declarative, serializable graph that separates authoring, compilation and execution into distinct layers.

• Agent frameworks

A work-in-progress Neovim plugin that uses a locally running Ollama model to suggest source code completions, similar to GitHub Copilot.

• Editors

A Rust-based local AI runtime and orchestrator that synchronizes text, vision and tool execution in one memory space, with hybrid HTTP and native FFI access for building agentic workflows on edge hardware.

• Agent frameworks
• Inference