FINGER TRADER

Python & SDK

Write a Python strategy

Create a code project, edit parameters and use the same native backtest execution path.

Updated 18 November 2026

Create a code project

  1. In My Strategies create a new strategy and choose Code / Python. This is a distinct authoring mode from the visual graph’s JSON code page.
  2. Open the generated strategy.py entry. It derives from finger_trader.hosted.Strategy. Use the generated project as the API-compatible starting point.
  3. Edit supported source files in Desk or export the source project to your preferred editor. The finger-strategy.json manifest defines parameter IDs, types, values and optional ranges.

Read native indicator results

Context.indicator(id, kind, parameters=...) requests FQE indicators. Use a stable independent instance ID and explicit calculation parameters. Numeric results are Decimal; an invalid warm-up result is None. Handle None before comparing values. Keep financial decimals exact rather than converting them through binary floats. Refer to the installed SDK’s types and generated project for callback signatures.

Validate, bind and backtest

  1. Run Check Strategy to validate the project manifest and Python syntax. Syntax validation does not execute imports or prove trading behavior.
  2. Bind a validated Python / SDK environment. Save the project.
  3. Open Replay Settings, choose local history and a simulated account, and set capital, sizing, leverage, protection and fees. Press Play after saving.
  4. Inspect Python logs, native fills and account snapshots together. Pause, step and stop control this execution; reopening a saved run does not execute Python again.

Portability and supported scope

Source import/export does not install requirements or copy .venv and .git. A portable Python model requires a compatible client and a new local environment binding after import. Current hosted live support has narrower order and instrument capabilities than arbitrary Python code: confirm support before deploying. Other strategy languages are not part of this first Python workflow.

For the current Desktop workflow. Labels and available capabilities can vary by application version and connected provider.