Search

Smart caching system for LLM responses - reduce API costs by avoiding redundant calls

A variable n-gram powered search engine for intelligent completion and semantic search

Vercel AI SDK adapter that exposes discoverActions and callAction.

Simple one-to-one WebRTC video/voice and data channels

Universal AI manifest, rules, skills, prompts, slash commands, and personas for the commerce-atoms ecosystem.

Build and maintain persistent LLM knowledge bases from any content.

A dead-simple Lavalink alternative powered by FFMPEG + yt-dlp.

Automatic memory context injection for pi coding agent. Load, search, and persist knowledge across sessions using Markdown packs.

Evolutionary multi-agent framework. Diverse agent personalities compete and evolve across generations to solve your task.

Low-token handoff workflow for multi-agent repos. Installs Claude Code slash commands and shared memory files.

AI coding support CLI tool - context management, file watching, and multi-agent review

Portable Obsidian vault bootstrap CLI for coding projects

Persistent template-backed tools for pi

Native Ollama provider for pi coding agent — fixes tool calling under streaming

Native Ollama model discovery for OpenCode.

AI-powered vulnerability scanner for any codebase

Agent-operated scaffolding compiler for generating standalone SolidStart apps.

Project-aware WCAG 2.1/2.2 auditor with AI-ready fix prompts for vibe-coding tools (Cursor, Claude Code, Windsurf).

Agent-ready music CLI — JSON output, batch pipe, zero config

MySQL state adapter for Chat SDK