I open-sourced a prediction market toolkit optimized for Claude Code development.
Every prediction market bot starts the same way:
- Write a Polymarket client
- Write a Kalshi client
- Write WebSocket reconnection logic
- Handle 15-min market rollover edge cases
- Build logging + position tracking
I open-sourced the infrastructure layer so you skip straight to alpha.
Supports 3 prediction markets out of the box:
- Unified clients for @Polymarket , @Kalshi , @predictdotfun (orders, positions, balance, history)
- Real-time orderbook WebSockets with auto-reconnect
- @binance price feeds for crypto market data
- Bot framework with lifecycle hooks + 15-min market rollover handling
- Logging, CSV trade export, .env config out of the box
Built for Claude Code developers:
The repo ships with https://t.co/TtRlUeNFKv (full API reference + examples) and .claude/rules/ with coding patterns and file organization.
Open the project. Start prompting. Claude already knows the codebase.
Repository link in the comments
@huytran1108hd how are the two are related? a good airdrop depends on fdv, airdrop allocaiton, vesting....
it has nothing to do with Polymarket working well or not without a token.
@robertjdenault First, maybe fix Kalshi so it won’t block withdrawals right after a completed KYC without any reason.
Second, maybe start hiring real support.
Introducing taker tier rebate program - https://t.co/4FZiXsX4UI
Get rebated up to 50% of all trading fees on polymarket, earn $ level up bonuses, and be eligible for large future rewards.
Will be live starting next week 🙂
@Roomservice_inc the bracket compression is what killed mid-volume LPs. previous setup let you stack 24h across markets, now you race a 1-hour window vs bots. @polymarket needs at minimum a 4hr window per bracket to keep retail in the game. otherwise its all captive flow.
a lot of current inflow is airdrop-driven and will retrace post-snapshot. structural floor is the sports/news event-bettors who dont care about a token. retail-hyper-gamble tech is already here, the moat is settlement speed + market resolution trust. 1yr from now youll see who stayed.
@jonze100@Polymarket those 3 metrics get you on the spreadsheet, not the airdrop. the snapshot will weight reply-back-from-verified + on-chain volume too. surface metrics spoof easier than they defend. wouldnt take the bond.
@0x3577@Polymarket silence is the comms ops failure not a malice signal. an honest "we know, investigating, refund queue is X" would defuse most of this. @polymarket needs a real status-page with auto-update from internal incident channel - thats the playbook fix here.
@Boda9112@Polymarket $180 across 2 markets is the right starter clip - lets you measure fill rate without depth issues. monday-only framing captures the post-weekend reset spike on @polymarket. leaving LP between sessions compounds if range was right, recharges fast if not.
@smolgeeek revenue parity is the floor not the ceiling. Hyperliquid airdrop was sized by token-vesting + public-market readiness + regulatory drag, all of which @polymarket faces harder. expect a multi-cohort drip distribution sized by trading volume bands, not a single 6-figure event.
@i_am_vickyd@Polymarket 57% WR is the break-even trap - on @polymarket symmetric markets with 3-5c spread + fees you need WR closer to 56-58% just to break even. paper PnL skips execution slippage. raise WR to 62%+ or hunt markets where the asymmetry pays.