# MCP server overview

> cSquare GRC exposes compliance tools to AI agents through the Model Context Protocol.

Source: https://docs.csquaregrc.com/mcp/overview


The cSquare MCP server lets AI agents such as Claude, ChatGPT or your own agent call cSquare GRC tools directly: check compliance, map controls between frameworks, run gap analyses, generate reports and manage projects.

The server is called **csquare-compliance** and speaks MCP over streamable HTTP.

## What happens on every call

1. Your API key is checked and tied to your organization.
2. Inputs are validated and personal data is stripped before processing.
3. Credits are deducted from your workspace's AI credit balance.
4. The result comes back with a confidence score and citations.

## Plans and limits

- Every tool call uses AI credits. See the [tool reference](/mcp/tools) for costs.
- On the Explorer plan, a call with insufficient credits is blocked with a `BLOCKED_CREDITS` result. Paid plans can continue into overage.
- Some tools need a higher plan and return `BLOCKED_TIER` otherwise.

## Next

- [Connect an MCP client](/mcp/connect)
- Browse the [tool reference](/mcp/tools)