Automation
Delivery modes, duplicate detection, and building a hands-off document pipeline.
Manual review is the default, but every step of the pipeline can run automatically once you trust a parser.
Delivery modes
Each destination on a parser has its own delivery mode:
| Mode | Behavior |
|---|---|
| Manual | Data is sent only when you click Approve & Send. The default. |
| Auto-send on completion | Data is sent as soon as extraction completes, with no review step. |
| Auto-send on approval | Data is sent automatically the moment a document is approved. |
Because the mode is per destination, you can mix strategies: for example, auto-send every completed document to a webhook for logging while keeping QuickBooks on manual approval.
There is no confidence-based auto-approval. Confidence scores are shown to guide your review, and delivery is controlled only by the modes above.
Duplicate detection
Parsedit hashes every incoming file. If a file with the same content hash already exists on the same parser, the upload is recognized as a duplicate:
- No new document is created.
- No pages are deducted from your quota.
- Destinations do not receive a second row.
This makes automated intake safe. An email forwarded twice or a Zapier retry cannot produce double entries in your spreadsheet or accounting system.
Automated intake
Combine automatic intake with an auto-send mode and documents flow end to end without touching the dashboard:
- Documents arrive via email forwarding, inbound webhook, Google Drive, Zapier, or the API.
- Extraction runs automatically, including OCR for scans.
- With auto-send on completion, results land in your destination immediately. With manual or auto-send on approval, documents wait in the review queue.
To stay informed without watching the dashboard, connect Slack notifications for received documents, completed extractions, failures, and delivery results.