Capability · Classify & Split
Upload a multi-page PDF of mixed documents. Parsedit detects each document type, splits the packet into separate files, and routes every segment for review.
Packets often mix invoices, receipts, IDs, and statements. Classification finds the boundaries; split turns each segment into its own document.
Detected document types and page ranges before split.
Classified packet
Intake often arrives as one scan: an invoice, a purchase order, and a receipt stapled together in a multi-page PDF. Manually separating those pages slows accounts payable, mailroom, and onboarding workflows before extraction even starts.
Parsedit’s document classify and split capability detects document types across page ranges—invoice, receipt, identity document, bank statement, purchase order, and related business forms—then splits the packet into child documents you can review and route independently.
Unlike splitting a PDF by fixed page counts, classification finds logical boundaries. Each segment keeps its original page range, so reviewers see exactly which pages belong to which document type without re-uploading or re-scanning.
Scanned packets work the same way as digital files. When embedded text is missing, OCR runs first so type detection and field extraction still apply—one pipeline for native PDFs and image-based multi-document uploads.
After the split, assign each classified segment to the right parser, or create a parser from the extract. Set a default parser per document type so the next similar packet routes automatically and stays out of the wrong schema.
Ambiguous pages stay available for manual assignment with classification confidence visible, so high-volume intake does not force a binary auto-or-fail choice. Teams keep control where layouts are messy or types overlap.
Once documents are typed and separated, the same extract–review–deliver flow applies: structured fields land in Google Sheets, webhooks, or your ERP-connected stack, with human approval when you want it—ideal for AP packets, KYC bundles, and mixed mailroom PDFs.
Purpose-built capabilities for this document type.
Identify invoices, receipts, IDs, statements, and more inside a single upload.
Split at document edges so each child file covers the right page range.
See how sure classification is so ambiguous packets get a human look.
One packet in. Typed, reviewable documents out.
Drop a multi-page PDF that contains more than one document type.
Intake often arrives as one scan: invoice, PO, and receipt stapled together. Classification labels each range so nothing is processed under the wrong schema.
One click turns the packet into child documents with page ranges preserved. Review each file on its own, or assign parsers by type for the next similar packet.
Once documents are typed and separated, the same schema-driven extraction and delivery pipeline applies—Sheets, webhooks, and review gates included.
See field extractionPre-built schemas for common document types. Customize fields and connect a destination.
Integrations
After review, rows flow to Sheets, webhooks, Slack, Airtable, Xero, and more.
Your documents stay in your account. Review is on by default, and auto-send only runs where you enable it.
No training on your data
Review on by default
Secure storage of extracted data
You stay in control
Common questions about this capability. Need more detail? our documentation
A packet is a multi-page file that contains more than one logical document—for example an invoice followed by a receipt in the same PDF.
Common types such as invoices, receipts, identity documents, bank statements, purchase orders, and related business forms. Ambiguous segments stay available for manual assignment.
No. You can review the packet as uploaded, or split into documents when you want separate review and routing per type.
Each segment becomes a child document with its page range. You assign a parser (or create one), then extract and review fields as usual.
Yes. When you assign a parser to a classified type, you can set it as the default for that document type on future packets.
Capability · Classify & Split
Create your first parser in minutes. No code, no setup calls.
Assign each classified segment to an existing parser, or create one from the extract.
After the split, every document has its own review surface—no more scrolling a mixed packet.
Parsedit detects document types and marks where each segment starts and ends.
Each segment becomes its own document, ready for field extraction and approval.