# ProjectFlow MCP Server

This is the official [Model Context Protocol (MCP)](https://modelcontextprotocol.io/) server for ProjectFlow. It allows AI coding agents (like Claude Desktop, Antigravity, Cursor, etc.) to securely read, update, and create tasks on your ProjectFlow boards.

## Installation

Any user can install this globally using pip:

```bash
pip install projectflow-mcp
```

## Configuration (For Users)

Once installed, users just need to add this server to their AI agent's configuration file.

### For Claude Desktop & Antigravity

Add the following to your `claude_desktop_config.json` (or `mcp_config.json` for Antigravity):

```json
{
  "mcpServers": {
    "projectflow": {
      "command": "projectflow-mcp",
      "env": {
        "PROJECTFLOW_EMAIL": "your-email@example.com",
        "PROJECTFLOW_PASSWORD": "your-password",
        "PROJECTFLOW_API_URL": "https://projectflow.thuthuka.tech/api"
      }
    }
  }
}
```

*Note: The `PROJECTFLOW_API_URL` defaults to the main production URL if not provided.*

## Available AI Tools
When connected, your AI gains the following capabilities:
* `list_projects`: Discover available workspaces and projects.
* `list_my_tasks`: View assigned tasks and current board status.
* `get_task_details`: Fetch subtasks and deep descriptions.
* `update_task_status`: Move tasks to `TODO`, `IN_PROGRESS`, or `DONE`.
* `create_task` / `create_subtask`: Allow the AI to log bugs or break down its own work.