Skip to main content
Before running scrapers, editing SQL models, or making any pipeline changes, create a workspace. All write operations require a workspace_id.

​
Create a workspace

1

Call workspace_manage with operation='create'

Pass a scraper_name to clone scraper data, or omit it for a SQL-only workspace.Scraper-based workspace — clones production data for a specific scraper:
SQL-only workspace — no data cloning, for model development:
2

Save the workspace_id from the response

The response tells you what was created:
For a SQL-only workspace:
Copy the ID — you’ll pass it as workspace_id to every subsequent tool call.
All pipeline write operations require workspace_id. This includes scraper_run, extraction_run, validation_run, warehouse_manage, sql_model_save, and any other tool that modifies data.

​
Clone from another workspace

Use from_workspace_id to branch off an existing workspace instead of starting from production. This copies the source workspace’s SQL models and scraper data into a new workspace.
This is useful when you want to iterate on work already in progress without modifying the original workspace.

​
List your workspaces

To see all active workspaces:
To filter by scraper:
Example output:
Promoted workspaces are shown with a [PROMOTED] label and cannot be deleted.

​
Delete a workspace

When you’re done with a workspace and don’t want to promote it, delete it to clean up:
You cannot delete a workspace that has already been promoted.