Import data and build a first graph
A short walkthrough of the Windows app interface, app source version 0.1.0 at revision 6acbdc33ce0a0af2440a797bd0e726b13e879572, reviewed 5 October 2026. The authentic screenshots were committed on 21 September 2026 and hash-tracked on 22 September 2026. Labels can change in later releases.
1. Index your own price files
Prerequisite (app source v0.1.0, reviewed 5 October 2026). Open the Windows app with a readable local CSV or Parquet file, or a folder containing them. Action. Use Data → Import, then Add file(s) or Add folder(s). Expected result. The Data dialog lists the selected path, instrument count and detected base timeframe, or a named scan error. Limit. Indexing lists files; price values are validated when loaded. Refresh scans after changing files on disk.
2. Connect a runnable graph
Prerequisite (app source v0.1.0, reviewed 5 October 2026). The scan has found the instrument and timeframe you intend to use. Action. Open Build, add a file data node, then connect Timeframe, Markets and a runnable Strategy. Select your instrument in Markets. Expected result. The canvas shows connected nodes and the selected market; the data node can use the imported folder when its path is empty. Limit. A catalogue entry marked as roadmap is not a runnable strategy. A run can still fail when a market is missing or an input is unsupported.
3. Run and inspect the result
Prerequisite (app source v0.1.0, reviewed 5 October 2026). A connected graph, imported bars and an available engine. Action. Select Run backtest in the Build toolbar, then open the resulting analysis. Expected result. A completed run exposes its recorded output, or a named refusal if preconditions fail. Limit. Preview is a short wiring check and returns no verdict. A completed backtest alone does not establish a valid trading edge.
Next: read validation and export coverage. For node-specific inputs use the node reference; for supported files and install requirements use installation.