Point BarZero at a folder of contracts — local disk, Google Drive, OneDrive, SharePoint — and review every one against the same playbook. Each contract still streams through the zero-retention path; the batch finishes with a cross-contract summary grouped by risk.
Each file is read by your browser, extracted to text on-device, and only the clause text is relayed to the model under ZDR. Tab-open is the default; close the tab and the batch pauses.
A queued worker fetches each file from the source you configured, sends it under ZDR, and discards the bytes after the response. No file is written to disk anywhere on BarZero infrastructure.
Pick a single playbook position for the batch. Each contract is reviewed against the same preferred clauses and walk-aways, so the report is comparable across the whole stack.
When the batch finishes, BarZero writes an aggregate index — every contract grouped by headline finding (red / amber / clean), top recurring risks, outliers worth a second look.
Failed files surface why — couldn't extract, file too large, password-protected. You'll know exactly which contracts to retry without scanning the whole queue.
Batch doesn't change the privacy posture. Each contract goes through the same single-doc inference path under Anthropic ZDR. No batch-specific telemetry, no retention exception.