Privacy
OpenCause Compute is built for volunteer processing of open-access/public research literature. The worker is not designed to process private medical records or personal documents.
What the worker sends back
Workers submit evidence candidates, summaries, validation warnings, node identifiers, worker version, platform/capability metadata, and model/extractor provenance needed for auditability.
Local files
The worker should use its app data directory for credentials and logs. It should not read personal files or documents outside its configured app data directory.
Operational records
The coordinator records events needed to run the system: enrollment, registration, heartbeat, work claims, submissions, validation, and admin actions.