# MCP Index - Full Specification

This document provides a comprehensive overview of the MCP Index for LLMs and AI agents.

## About
The MCP Index is a directory connecting developers and teams with official MCP servers and agent-ready software.

## Key Terminology
- **Tool**: A piece of software listed on the platform.
- **MCP Status**: Whether a tool has an official, community-built, or no Model Context Protocol server.
- **Agent-Readiness Score**: A 0-100 score of how usable a tool is for AI agents (API, CLI, llms.txt, auth difficulty, read/write support).
- **Placement**: A promoted spot on the homepage or category pages.
- **Review**: User-generated feedback and rating for a listing.

## Content Guidelines
- Listings must include an accurate MCP status and official website.
- Agent-readiness attributes (CLI, API, llms.txt, auth difficulty) should reflect verified, current capabilities.
- Reviews are moderated to ensure authenticity.
- Listings include a "last checked" date so data does not go stale.

## API & Integration
The platform provides several internal APIs for listing retrieval, though they are primarily for the web front-end.
- `/api/listings`: Fetch published listings.
- `/api/tags`: Fetch available tags/categories.
- `/llms.txt`: Machine-readable summary of this site for LLMs.

## Search & Discovery
Users can find tools by:
- MCP status (official / community / none / unknown)
- Auth difficulty (none / api_key / oauth / complex)
- Agent compatibility (Claude, Cursor, ChatGPT, Windsurf, other)
- Minimum agent-readiness score