Put a Lottery Data Warehouse Inside Claude, ChatGPT and Gemini
PatternSight ships an MCP server. Connect it once and your AI assistant can query 1.17M draws and run real analysis — without leaving the chat.
The most useful place to ask a data question is the tool you already have open. For a growing number of people that is an AI assistant — Claude, ChatGPT, or Gemini. So PatternSight ships a Model Context Protocol (MCP) server that plugs the entire lottery warehouse and analytics stack straight into those assistants.
What your assistant can do once it is connected
The server exposes eight tools: list_games and get_game for the catalog, latest_results for the most recent draws, frequency / gaps / analyze_game for descriptive analytics with chi-squared context, and predict / backtest for the prediction swarm with honest, walk-forward evaluation. Ask in plain language — "show me the hottest Mega Millions numbers over five years" — and the assistant calls the right tool and reads back the result.
Same data, same honesty
Everything the MCP server returns carries the same framing as the rest of PatternSight: descriptive analysis over real draws, with outcomes never guaranteed. The analytics cover all 286 games; the swarm runs picks and backtests on the games with sufficient baked history. It is the full warehouse — just delivered where you work.
Connecting takes one configuration entry. The MCP reference documents the transport, auth, and every tool.