Content classification is different from invoice-field extraction
A document can be recognized as an invoice without reliably identifying its supplier, issue date, invoice number or payment status. Foldora classifies against its folder categories and may suggest a short descriptive name. Names are summaries, not accounting records: dates and identifiers are not guaranteed to be retained. A strong filename match can finish classification before content is read.
Invoice-relevant text readers include PDF, DOCX, TXT, CSV and XLSX. PDF text is sampled from at most the first three pages; TXT reads up to 3,000 characters, DOCX up to 200 nonempty paragraph parts, and spreadsheets use a small row summary rather than the entire workbook. The text-classification path caps content at 5,000 characters; name generation can use a much shorter selected line. Do not assume every field or page is inspected.
Content analysis has an 8 MiB per-file size guard. For a PDF whose first page has fewer than 50 extracted characters, the reader attempts local OCR on that page at 150 dpi. This requires a working OCR runtime and can fail; it is not full-document OCR. A photograph of an invoice is not a supported structured invoice-extraction workflow. Unreadable text, scans and larger files need manual checking, and filename fallbacks may still suggest a category.
Foldora runs on Windows 10/11 and macOS. Document analysis is local; purchase, download, license activation, periodic license validation and updates require internet access. This guide is published by Computora AI, Foldora's publisher.
If you already know the fields your accounting workflow requires, draft a pattern with the free naming convention generator.
Comparing low-setup AI suggestions with Windows rules tools? Use the Windows file organizer comparison to choose the workflow that fits.
Frequently asked questions
Does Foldora extract the vendor, invoice date and invoice number?
Not as validated structured fields. Content classification and short filename suggestions may reflect document words, but there is no invoice-field schema or guarantee that vendor names, dates and invoice numbers are identified or preserved. Verify them against the original document.
Can I enforce exact vendor/date folders in Foldora?
Foldora has no configurable exact vendor/date routing rules. AI Content suggests categories; Date mode groups by filesystem last-modified year and month, not invoice issue date; Type mode groups file types. Choose another workflow if a prescribed accounting folder schema is mandatory.
Can Foldora analyze every page of a scanned invoice?
No. PDF text reading is limited to the first three pages, with a local OCR attempt on the first page only when its extracted text is short. OCR availability and readability are not guaranteed, and content analysis skips files over 8 MiB. Review scans and later-page details manually.
What must I review before moving invoice files?
Review Preview destinations and proposed names against each original invoice. Check required identifiers, ambiguous classifications and unreadable files; retain a backup. Foldora does not validate amounts, line items, payment status, tax treatment or reconciliation.