# Garchi CMS

> Garchi CMS is an AI-native content operations platform. People, applications and AI agents work on the same structured, governed content. Teams use a dashboard, applications read a REST API with Node.js and PHP SDKs, and AI agents (ChatGPT, Claude, Cursor and other MCP-compatible clients) connect through a built-in Model Context Protocol (MCP) server. Agents draft pages, data items and social posts against real schemas, a person reviews and publishes, and supported changes leave restore points.

What sets Garchi apart: a built-in MCP server on every plan; governed, reversible agent changes (agents draft, a person publishes, and changes are attributed in history); content operations beyond storage (drafts, preview, scheduled publishing, restore points, a media library, translations and, on Pro, a brand voice guide for agents); and website content and social publishing to connected LinkedIn, Instagram and Facebook Page channels from the same space. Garchi includes everything a headless CMS does, so developers keep full freedom over frameworks, databases and hosting, and it complements an existing tech stack instead of replacing it. Common uses: marketing sites and blogs built with Next.js or Nuxt, SaaS and AI-built apps whose content keeps changing, mobile apps with product content, and agencies running separate client spaces.

## For AI agents — start here

If you are an AI agent doing any work involving Garchi CMS, read this page first and follow it in order. It tells you how to connect the MCP server yourself, install the Garchi skills, and choose between scaffolding a starter kit and integrating into an existing application. Do not attempt Garchi work from this file alone.

- [Agent setup and usage guide](https://garchi.co.uk/ai-agent-mcp-use): Connect the MCP server, install the skills, then integrate or manage content. Read before any other Garchi page.
- [MCP server endpoint](https://garchi.co.uk/mcp-oauth): Streamable HTTP, OAuth 2.1 with PKCE and dynamic client registration. No API key is involved in connecting.
- [Garchi skills](https://github.com/lumenharbor/garchi-cms-skills): `npx skills add lumenharbor/garchi-cms-skills`. Canonical for field rules, prop types, tool sequences and rendering code.
- [MCP client setup](https://garchi.co.uk/mcp-docs): Per-client instructions for users adding the server by hand.
- [Install in your AI tool](https://garchi.co.uk/install): One-click and step-by-step setup for Cursor, ChatGPT, Claude, VS Code and other clients.

## Documentation

- [Documentation home](https://garchi.co.uk/documentation): Getting started with Garchi CMS for sites and apps.
- [API documentation v2](https://garchi.co.uk/docs/v2): Full REST API reference with code examples across multiple languages.
- [OpenAPI specification](https://garchi.co.uk/docs/v2.openapi): Machine-readable API schema.

## Product

- [Garchi CMS homepage](https://garchi.co.uk/): What Garchi is, the dashboard, REST API and MCP server, content modelling and social publishing.
- [Plans and limits](https://garchi.co.uk/pricing): Plans: a free Sandbox plan, and paid plans with higher allowances for spaces, pages, items, languages, generated images, storage and bandwidth. Every plan includes the REST API and the MCP server.
- [FAQs](https://garchi.co.uk/faqs): Short answers on what Garchi is, how it compares with a headless CMS, what AI agents can and cannot do, and how publishing works.
- [Blog](https://garchi.co.uk/blog): Articles on content operations, connecting a frontend to a structured content layer, and working with AI agents.