SignalAI SDK

Write a trading strategy in Python, backtest it honestly, and ship it.

You write a strategy in Python and backtest it in the cloud — signalai submit runs it on the SignalAI engine and returns the equity curve, trade ledger, and stats, viewable at your backtests.

Why "honest"

Backtests lie when they peek at the future or assume perfect fills. The SDK is built to avoid that: orders never fill on data they couldn't have seen (see Fills), and intraday strategies are evaluated the way real intraday trading works — each day from the same capital, flat overnight (see Intraday vs long).

Install

# the CLI (runs backtests in the cloud)
curl -fsSL https://signal-ai-mu.vercel.app/install.sh | sh

# the Python library (write strategies with autocomplete)
pip install signalai-quant

Then signalai login, scaffold a strategy with signalai init my_strategy.py, and signalai backtests submit it. Full walkthrough: Setup and the tutorial; or browse the SDK reference.

Three ways in: the web app, the CLI, or the Claude connector. See Ways to use it.