Use cases
What people build with polybot.
polybot is a framework, not a single bot — and not a bespoke script per venue. These are the shapes teams most often reach for as agents start trading prediction markets in 2026. Each maps to built-in strategies, the MCP server, and the same paper-trading-default risk invariants.
Cross-venue arbitrage
Capture price differences on the same event across Polymarket, Kalshi and Opinion — with a unified position model that nets exposure across venues.
Read the use case → AgentsAn AI agent trading desk
Give Claude or GPT an MCP-native trading desk: it analyses strategies, proposes changes and drafts orders that you approve at the CLI.
Read the use case → LiquidityMarket making on prediction markets
Run passive spread-farming strategies that quote both sides and earn the bid-ask, with inventory and exposure caps enforced by the platform.
Read the use case → ResearchStrategy research & shadow testing
Prototype a BaseStrategy, run it in shadow mode against live markets, and use the DuckDB analytics store to measure it before risking capital.
Read the use case →Need an agent system built like this?
Cryptuon builds production AI agents, MCP integrations, and trading systems. polybot is our open-source showcase.