The user has built a library for prediction market arbitrage and plans to add a fully automated bot that listens to the scanner and auto-executes on the spreads.
A few days ago, I posted [here](https://www.reddit.com/r/algotrading/comments/1q83w3d/found_5_arbitrage_spreads_in_prediction_markets/) about the arbitrage spreads I found between Polymarket and Kalshi. The response was great, but the consensus: finding the spread is easy, executing it before it closes is the hard part. The library is no longer just a scanner; it now supports native order execution. I’ve abstracted away the complexity, so you can now programmatically buy/sell positions on both platforms directly from the library: const client = new pmxt.Polymarket({ privateKey: your-key-here }); // or pmxt.Kalshi const order = await client.createOrder({ marketId: '663583', outcomeId: '109918...', side: 'buy', type: 'market', amount: 10 }); The goal is to move from "monitoring a dashboard" to "atomic-ish execution" where you can hit both legs of the arb almost simultaneously. Now that the execution primitives are done, the next update will be a fully automated bot example that listens to the scanner and auto-executes on the spreads. I'll be back in a few days with this update! [https://github.com/qoery-com/pmxt](https://github.com/qoery-com/pmxt) For those of you already trading these markets, are you finding that `market` orders are reliable enough given the lower liquidity, or do you strictly stick to `limit` orders to avoid slippage on the second leg?