Check extracted document data against its source with VCMP Unfold

Choose the output before uploading the document

Unfold accepts PDFs, Word documents, images, phone photos and plain text. Preset or custom fields help define what a reviewer needs from a document, while table extraction and Markdown support different output shapes. Start with one representative document and choose the fields that make its result useful.

Inspect the source excerpt beside each extracted value, review warnings and edit the result when needed. Missing details stay empty rather than being filled with invented information. A clear image and a defined field set are a better evaluation than a document whose expected answer nobody has agreed.

Review before exporting a batch

Batches can contain up to ten documents. A combined CSV uses one row per document; table CSV, JSON and Markdown exports support other uses. Re-uploading the same file with the same fields within 30 days can reuse the existing result without another document credit. Extraction results are available for 30 days.

Unfold does not automatically post extracted data into accounting software or other VCMP products. Treat an export as data your team still needs to check and place in the right workflow. Original uploads are processed in memory; their contents are sent to the AI provider for processing, where provider retention may apply. Confirm this handling fits the documents you plan to use.

Your evaluation checklist

1. Select representative documents and agree the fields or table structure needed.
2. Compare extracted values with their source excerpts and review missing fields.
3. Check column names and one-row-per-document behaviour in a combined CSV.
4. Confirm current file limits, result availability and document handling before a live batch.

Sources and next steps

Product facts reviewed 5 October 2026. The publication date above is the date of this note.

Check the official product page and buyer guide for current capabilities, scope and terms before choosing.