# FeelFit MCP Server

MCP (Model Context Protocol) server for accessing body composition data from FeelFit smart scales. Supports multiple accounts.

## Features

- Multi-account support
- Body composition measurements (weight, body fat, BMI, muscle mass, bone mass, water, protein, visceral fat, metabolic rate, body age)
- Device management (bound scales)
- Health goals
- Automatic authentication with token caching

## Tools

| Tool | Description |
|------|-------------|
| `list_accounts` | List configured accounts and auth status |
| `login` | Authenticate account(s) |
| `get_profile` | Get user profile |
| `get_measurements` | Get all body composition measurements |
| `get_latest_measurement` | Get most recent measurement |
| `get_devices` | List bound smart scales |
| `get_goals` | Get health goals |
| `get_all_data` | Get complete data dump |

## Setup

### 1. Install dependencies

```bash
pip install -r requirements.txt
```

### 2. Configure accounts

Create `~/.config/feelfit-mcp/config.json`:

```json
{
  "accounts": [
    {"email": "user@example.com", "password": "your_password"}
  ]
}
```

Or set the `FEELFIT_CONFIG` environment variable to a custom path.

### 3. Add to Claude Desktop / Claude Code

```json
{
  "mcpServers": {
    "feelfit": {
      "command": "python3",
      "args": ["/path/to/mcp-feelfit/server.py"]
    }
  }
}
```

## How it works

The server connects to the FeelFit Cloud API (`feelfit.qnclouds.com`) using the same protocol as the Android app. Passwords are encrypted with RSA before transmission. Authentication tokens are valid for 180 days.

## License

MIT