This tab contributes compute to a distributed query session. Keep it open while you want to volunteer resources.
Select how you want to handle incoming query chunks before this worker connects to the coordinator.
The coordinator wants to run the SQL below locally in your browser against a remote dataset. Review it, then choose Execute or Decline. Declining skips all chunks for this run; you will be prompted again for the next run.
This is an open-source experiment. Bugs should be expected.
How it works: A researcher creates a session and shares this worker URL with volunteers like you. By keeping this tab open, your browser receives chunks of a dataset from the coordinator, processes each chunk locally using DuckDB-WASM, and returns partial results — without sending raw data to any server. The coordinator merges all partial results once all chunks are complete.
Use Auto mode to process chunks as they arrive, or Confirm mode to review the SQL before each run executes.