AI-powered trade journaling and pattern analysis for futures traders.
pnpm install
cp .env.example .env.local # fill in Tradovate + Anthropic + Supabase
pnpm daily # runs on today's tradesEvery day, this pipeline:
- Pulls your filled executions from Tradovate
- Pairs entries with exits to reconstruct trades
- Computes R-multiple, P&L, and commission per trade
- Asks Claude to classify each trade: setup, time-of-day, quality, mistakes, notes
- Generates a Danish-language daily summary in markdown
- Saves the summary to your brain folder and optionally pings you on Telegram
Most trade journaling tools are static. You log a trade, the tool stores it. That's it.
This one reads your trades. It tells you that you lose most on pre-market shorts that reverse at support. It tells you that your A+ setups happen between 14:30-15:00 UTC. It tells you that "moved stop" appears in 60% of your losing trades.
The LLM does the pattern-finding so you can spend your time trading instead of building spreadsheets.
- Tradovate account (live or demo)
- Anthropic API key
- Supabase project (used for trade + classification storage)
- Node 20+ and pnpm
See docs/setup.md.
See docs/architecture.md.
Alpha — built by Noah Duus (prop trader). Used daily on my own account. Public repo as part of Phase D of the 90-day AI Automation Business launch.
MIT. Use it, fork it, modify it. If you commercialize it, a credit would be appreciated but not required.