CLI

signalai is the command line for the SDK. Install it with:

curl -fsSL https://signal-ai-mu.vercel.app/install.sh | sh

init <path>

Scaffold a new strategy file.

signalai init my_strategy.py

login

Sign in so submit / status can reach your account. signalai login opens your browser, authenticates, and stores the credential locally (macOS Keychain; an 0600 file on Linux).

signalai login

Headless / CI (no browser): store a token directly with signalai login --token <your-token> (from API tokens), or set SIGNALAI_TOKEN in the environment.

check <file>

Verify a strategy file builds on your machine — it imports the file, instantiates your Strategy, and calls init(), catching syntax, import, name, and init() errors before you upload anything.

signalai check my_strategy.py

Runs in your local Python, so it needs pip install signalai-quant (plus whatever your strategy imports). signalai backtests submit runs the same build check first and won't send a strategy that doesn't build. Library users can call it directly: from signalai_quant import check_strategy_builds → check_strategy_builds(path) returns a list of problems (empty when valid).

Commands

Commands are grouped by noun. Every read command accepts --json.

signals

signalai signals list
signalai signals show <signal_id>
signalai signals history <signal_id> [--from 2024-01-01] [--to 2024-06-01] [--limit 500]

backtests

signalai backtests submit <file> --symbols AAPL,MSFT --from 2015-01-01 --to 2024-12-31 [--resolution daily|minute]
signalai backtests list
signalai backtests get <backtest_id>

View full backtest results at your backtests.

strategies

signalai strategies list
signalai strategies show <strategy_id>

signalai submit and signalai status still work but are deprecated — use signalai backtests submit / signalai backtests get.