How to Get Prediction Market Data in Python in 60 Seconds

#python #api #data #tutorial

Ever wanted to know what prediction markets think about the World Cup, the midterm elections, or tomorrow's NBA games?

Here's how to get that data in 60 seconds.

Install

pip install meridianedge

Get consensus data

from meridianedge import MeridianEdge

me = MeridianEdge(api_key="your-key")

# What do markets think about upcoming events?
markets = me.markets(category="sports", limit=5)
for m in markets:
    print(f"{m['name']}: {m['consensus']}%")

# What shifted the most today?
movers = me.movers(hours=24)
for m in movers[:3]:
    print(f"{m['name']}: {m['shift']}% shift")

What you get

Real-time consensus probabilities aggregated from multiple regulated prediction markets. One API, one number per event.

27,000+ active markets tracked
Politics, sports, economics, global events
14-second data pipeline updates
Cross-platform divergence detection

Also available as

Node.js: npm install meridian-edge
MCP Server: pip install meridian-edge-mcp (works with Claude Desktop, Cursor, VS Code)
REST API: meridianedge.io/docs.html

Why consensus beats single-source

A single prediction market can be wrong. Consensus across multiple platforms is more reliable — just like how aggregating polls beats any single poll.

When platforms disagree (divergence), it often signals new information entering the market. That's where the interesting data lives.

MCP server for AI agents

If you're building AI agents that need to understand probabilities of real-world events, the MCP server plugs directly into Claude Desktop, Cursor, or VS Code:

pip install meridian-edge-mcp

Add to your Claude Desktop config:

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