Prediction Market Data for AI Agents

Why AI Agents Need Prediction Market Data

Language models are trained on static data with a knowledge cutoff. Prediction markets provide something different: real-time, crowd-sourced probability estimates that reflect what the market believes right now. When an agent can query "What probability does the market assign to X?", it grounds its reasoning in live data rather than stale training data.

This is particularly valuable for agents that make decisions under uncertainty, generate research reports, or answer questions about current events.

MCP Server (Model Context Protocol)

The Meridian Edge MCP server works with Claude Desktop, Cursor, VS Code, and any MCP-compatible client. Install with one config block:

{
  "mcpServers": {
    "meridian-edge": {
      "command": "uvx",
      "args": ["meridian-edge-mcp"],
      "env": {"MERIDIAN_EDGE_API_KEY": "your-api-key"}
    }
  }
}

Five tools are exposed: get_consensus, get_opportunities, get_signals, get_markets, and get_settlements. The server is listed in the Official MCP Registry.

Python SDK for Agent Frameworks

The Python SDK integrates directly with LangChain, LlamaIndex, CrewAI, and any agent framework that can call Python functions:

from meridianedge import MeridianEdge
client = MeridianEdge(api_key="your-key")
events = client.consensus(vertical="sports")

Node.js SDK

The Node.js SDK works with Vercel AI SDK, LangChain.js, and any JavaScript agent framework. Zero dependencies, native fetch.

A2A Protocol

For agent-to-agent communication, Meridian Edge supports the A2A protocol. The agent card is published at /.well-known/agent-card.json, enabling automatic discovery by other agents.

What Agents Can Do

• Answer questions: "What does the market say about tonight's game?" — grounded in live data
• Monitor events: Track probability shifts and alert when thresholds are crossed
• Research reports: Generate analysis that includes current market-implied probabilities
• Detect divergence: Flag events where platforms disagree significantly