An MCP server that connects Claude Desktop to Rhyme's live Supabase database. Ask Claude natural language questions about your platform data.

Setup

1. Install dependencies

pip install mcp supabase python-dotenv

2. Configure credentials

cp .env.example .env

Edit .env and add your Supabase URL and service role key

3. Run the simulator (optional — adds realistic demo data)

python simulate.py

4. Connect to Claude Desktop

Add to ~/.config/Claude/claude_desktop_config.json:

{
  "mcpServers": {
    "poetryquill": {
      "command": "python",
      "args": ["/full/path/to/server.py"]
    }
  }
}

5. Restart Claude Desktop

Look for the hammer icon — your tools are connected.

Example questions to ask Claude

• "Give me a platform overview of PoetryQuill"
• "What are the top 10 most liked poems?"
• "Who are the most followed authors?"
• "Analyze engagement trends over the last 30 days"
• "What tags are writers using most?"
• "Show me the user growth over the last 90 days"
• "Which collections have the most poems?"
• "Analyze a sample of recent poems and tell me what themes are trending"
• "Who has the longest writing streak?"
• "Where are our users from?"

Tools available

Tool	What it does
get_platform_overview	Totals, ratios, 30-day growth
get_top_poems	Most liked/viewed/commented poems
get_top_authors	Most followed/active/points authors
get_engagement_trends	Activity over a time window
get_content_sample	Poems for Claude to analyze themes
get_user_growth	Weekly signup trends
get_tag_analytics	Tag usage and engagement by tag
get_streak_leaderboard	Current and longest writing streaks
get_collections_overview	Public collections and sizes
get_geographic_distribution	Where users are from

Schema

Built for Rhyme's Supabase schema: profiles, poems, likes, comments, follows, collections, collection_poems