Publishing
Save a selected experiment revision as an app, with explicit control over creation and updates.
Publishing saves a selected revision into an app. It does not start a sandbox, deploy a strategy, or place live orders. Launch the published app separately through the normal app workflow.
Select the exact revision
Save pending editor changes, open History, and select the revision you intend to publish. Check the identifier shown under Publish revision against the run you reviewed.
Publication uses that selected snapshot. Selecting an older run does not automatically select its revision for publication.
Create an app
Choose Create app to create an app from the selected revision. When publication completes, choose Open published app and review the saved configuration.
The experiment's chat, annotations, and run history remain research context in the Lab. Publishing is not a transfer of an active historical execution into a live session.
Update an imported app
Experiments created by import also offer Update imported app. This targets the saved app from which the experiment was created.
The Lab checks that the imported app has not changed since its snapshot was taken. If someone edited it elsewhere, publication fails with a conflict instead of overwriting those changes. Import its current saved version into a fresh experiment and reconcile the changes, or use Create app to preserve a separate candidate.
After a successful update, further updates require a fresh import. The experiment is not a continuing synchronization link to the saved app.
Review before launch
Open the published app and verify its source, instrument, sizing, data configuration, and intended execution mode. Historical runs use simulated execution and do not prove broker-specific or live behavior.
Continue with backtest and launch when you are ready to configure a session.