> ## Documentation Index
> Fetch the complete documentation index at: https://docs.make-pretty.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Pretty MCP

> Use Pretty chats from Codex, Claude Code, and other MCP clients.

[Model Context Protocol (MCP)](https://modelcontextprotocol.io/) lets AI clients discover and call tools from another service. Pretty runs an MCP server for its chats. Connect Codex, Claude Code, or another client, sign in with your Pretty account, and ask it to use Pretty's tools. Chats started this way also appear in the [web editor](https://editor.make-pretty.com).

The connection lets your client start and continue chats, check when Pretty has finished, read replies, browse files already in your Pretty Workspace, and read available skills. Pretty runs the work in the workspace you choose when you authorize the client.

## Connect

Use a client that supports remote MCP over Streamable HTTP and OAuth. The server URL is:

```text theme={null}
https://api.make-pretty.com/mcp
```

Your Pretty workspace must have external access enabled. If authorization says it cannot be reached from outside the app, ask your workspace admin to [contact support](/help/support).

Follow [Connect an AI client](/mcp/connect) for Codex, Claude Code, and other MCP clients.

## How a chat works

Your client calls `start_chat` with a message. Pretty returns a chat ID immediately and works on the request. The client uses `get_chat_status` to check progress and read the latest reply. If Pretty asks a question, the client sends your answer with `send_message` using the same chat ID.

The [tool reference](/mcp/tools) covers every available tool, message attachments, chat statuses, and pagination.