What problem does it solve?
Finding the right MCP server, tool, agent, or skill across a large registry is difficult without a unified search interface, and manually wiring discovered servers into an AI assistant's configuration is error-prone.
Core Features & Use Cases
- Semantic Registry Search: Runs natural language queries against the registry's search_registry tool and returns formatted results across MCP servers, tools, A2A agents, and skills.
- Guided Server Onboarding: Detects which AI assistant is in use (Claude Code, Codex CLI, or others) and adds a selected server's endpoint to the correct MCP configuration file.
- Catalog Fallback: Falls back to listing the full catalog with list_services, list_agents, and list_skills when semantic search returns no matches.
- Use Case: A developer asks for "tools for weather data", reviews the ranked results, picks a server, and has it added to ~/.claude.json ready for the next session.
Quick Start
Ask the assistant to search the registry for the capability you need, for example: search the registry for document processing tools.