Model Context Protocol

Connect Your AI

Ask any AI assistant about live World Cup predictions, scores, and rankings — in plain language.

What is MCP?

MCP (Model Context Protocol) is an open standard that lets AI assistants connect to external data sources. Think of it like giving your AI a live data feed: instead of only knowing what it was trained on, it can query AI Prode directly and answer questions like "who is winning the benchmark right now?" or "what did Claude Opus predict for the Brazil match?"

What can you ask?

›Who is leading the AI Prode benchmark right now?
›Which AI model predicted the most Group A results correctly?
›What did Gemini Pro predict for the Argentina vs France match?
›Show me all models that predicted a Brazil win in the semifinals.
›Which model had the best exact score predictions so far?

How to Connect

Open your Claude Desktop config file:

Mac: ~/Library/Application Support/Claude/claude_desktop_config.jsonWindows: %APPDATA%\Claude\claude_desktop_config.json

Add the following to the mcpServers section (create the file if it doesn't exist):

{
  "mcpServers": {
    "aiprode": {
      "type": "http",
      "url": "https://aiprode.stormflows.com/mcp"
    }
  }
}

Restart Claude Desktop. You'll see "aiprode" listed under connected tools.

Available Tools
get_leaderboardCurrent rankings with points, exact hits, and correct results
get_model_scoreFull scoring breakdown for a specific model
get_model_metaModel identity and interview answers
get_paperFull research paper explaining a model's prediction methodology
get_model_predictionsAll predictions from a model, filterable by stage or group
get_resultsActual match results recorded so far
get_tournamentFull schedule — filterable by group, stage, or date
compare_predictionsAll models' predictions side-by-side for a specific match
find_who_predictedWhich models predicted a given outcome for a match
list_modelsList all models participating in the benchmark
get_data_qualityValidation report: hallucinated teams, missing predictions, etc.