calesthio/OptionsCanvas

Chart-native options trading you run on your own machine.

Drag your entries, stops, and take-profit targets directly on the underlying's chart. Stops never touch the broker order book until they're actually breached — HFTs can't hunt what they can't see. Everything runs on localhost; zero telemetry, zero SaaS.

Highlights of this first release

Trade from the chart, not from a ticket

• Drag the Buy, SL, and TP pills directly on the underlying chart — the price line is the control
• Drag Buy off live price → flips into a limit-trigger with an anchored ENTRY line
• One-gesture bracket orders (Shift+B): click entry, drag to TP, release
• TradingView-style hotkeys: B / S / F (flatten all) / 1-5 (size presets) / Alt+1..6 (timeframes)

Stops your broker can't see

• SL and TP live in the local engine + SQLite, NOT in the broker order book
• Only at the moment the underlying actually breaches your level do we send a market order
• No resting stops for market-makers to sweep on a wick

Think in the underlying, trade in options

• Set SL/TP in underlying dollars (not option premium)
• Black-Scholes overlay projects option premium and your P&L while you drag
• Auto-picks the contract: best DTE from your config, ATM strike from the live chain, both fresh from the broker

First-run browser wizard

• Pick broker → paste keys → "Test connection" → onboard a universe → "Start trading"
• No terminal, no JSON editing, no manual setup
• Recommended 30 highest-flow names by default; full 110-name universe on a single toggle; or custom watchlist

Automatic trade journaling

• Every fill, every close, every realized P&L written to local SQLite as it happens
• Daily aggregates (win/loss/net P&L) materialized on every trade
• JSON snapshots in assisted_trading/journal/<date>/trades.json for human review
• Query the DB with pandas.read_sql, DuckDB, Datasette, Jupyter — your data, your tools

Multi-broker abstraction

• Alpaca — production-tested (paper + live)
• Tradier implementation is in the repo but not yet end-to-end validated; see Broker support status
• Adding a new broker: one file in backend/, one entry in broker_registry.py. The wizard UI renders dynamically from the registry — zero frontend edits needed

One-line install

• Windows: double-click OptionsCanvas.bat
• macOS / Linux: ./optionscanvas.sh
• Docker (any OS): docker compose up -d
• On first run: creates a venv, installs deps (~2 min). Subsequent launches are instant.

Safety belt for live mode

• Paper trading is the default on every supported broker
• Switch to Live in the wizard → terminal banner + server log warning + red broker pill + app-wide red border + [LIVE] browser-tab prefix
• Impossible to end up in live mode by accident

Getting started

# Option 1: native (recommended)
git clone https://github.com/calesthio/OptionsCanvas.git
cd OptionsCanvas
./optionscanvas.sh   # or OptionsCanvas.bat on Windows

# Option 2: Docker
git clone https://github.com/calesthio/OptionsCanvas.git
cd OptionsCanvas
docker compose up -d

# Then open http://localhost:5001/setup and follow the wizard.

What's NOT in v0.1.0 (transparency)

• Tradier: code is there, end-to-end testing in progress — flip to officially supported in a future release once validated
• IBKR / Tastytrade / Schwab: the abstraction makes these mechanical to add (one file + one registry entry), but not implemented yet
• Multi-leg strategies (spreads, condors): single-leg only for now
• Mobile-responsive UI: desktop-first; mobile is best-effort

License

AGPL-3.0-or-later. Run a modified version as a network service → you must share the modified source. Fork it privately for your own trading → you're fine.

Disclaimer

Paper trading is the default. Options trading involves substantial risk of loss. Not financial advice. No warranty. If you don't understand exactly what an order will do before you place it, don't place it.