Polymarket Paper Trader

Install polymarket-paper-trader via pip. Python 3.10+ is required. Alternatively use the ClawHub installer for OpenClaw agent environments.

pip install polymarket-paper-trader

Run the CLI to create a paper account with an initial simulated balance. The default data directory is used unless you specify PM_TRADER_DATA_DIR.

pm-trader init --balance 10000

Add the server to your MCP client configuration file. The command pm-trader-mcp is installed automatically with the package and requires no API keys.

{
  "mcpServers": {
    "polymarket-paper-trader": {
      "command": "pm-trader-mcp"
    }
  }
}

Two optional environment variables control data storage and the active account. Set PM_TRADER_DATA_DIR to a custom path if you want to store account data outside the default location, and PM_TRADER_ACCOUNT to switch between multiple named accounts.

export PM_TRADER_DATA_DIR="/path/to/data"
export PM_TRADER_ACCOUNT="my-account"

Use the CLI or via the MCP tools to search for prediction markets and place simulated buy orders. The order book execution mirrors Polymarket's real fee formula and processes orders level by level.

pm-trader markets search "bitcoin"
pm-trader buy will-bitcoin-hit-100k yes 500
pm-trader stats --card

Run one of the included example strategies against historical data to benchmark your trading logic before using it with live paper trades.

pm-trader benchmark run examples.momentum.run
pm-trader benchmark run examples.mean_reversion.run