Document marked non-financial
Recover an invoice that the classifier filtered out before it reached Documents.
Document marked non-financial
An email arrived with a PDF that should have become an invoice or receipt, but it never showed up in Documents.
Likely cause: pre-classification skipped it
Every PDF that arrives through an email source is checked by an AI classifier before anything is stored. The classifier labels each file financial or non-financial. When it labels a file non-financial with a confidence of 0.7 or higher, Inbox Ledger skips that file completely. Nothing is written to storage, no extraction runs, and no credit is spent.
The skip keeps your document list clear of marketing PDFs, signed contracts, and newsletters. The cost is the occasional false positive, where a real invoice gets filtered out.
The skip applies only to email sources. Manual uploads and forwarding addresses are never pre-classified, so they always reach extraction.
What tends to get misclassified
- Faded or low-quality scans, such as handwritten receipts or phone photos of paper.
- Invoices in less common scripts. Some Arabic, Hebrew, and Asian-script layouts are still being tuned.
- "Order confirmation" emails from vendors that read like marketing.
- Hybrid PDFs, where one invoice page is followed by five pages of terms and conditions.
Exact fix: re-ingest through a bypass path
The skipped file was never saved, so there is no button in the dashboard to bring it back. Send the same PDF in through a path that skips the classifier.
Forward the email
Forward the original email to one of your forwarding addresses, which look like {code}@fw.inboxledger.app.
Forwarded mail goes straight to extraction, the same as a manual upload.
See Forwarding addresses to create one.
Upload the PDF by hand
Save the PDF to your computer, open Dashboard → Documents, and drag the file onto the upload area. Manual uploads always run the full extraction pipeline.
Report the false positive
Email support@inboxledger.app with the PDF attached. These reports feed back into the classifier and reduce the same misses over time.
Confirm it worked
Open Documents and watch for the file. After a forward or manual upload, it appears as a new row and starts extracting within a minute. Once the vendor and total are filled in, the document is in and counts as one credit.
Related
Pre-classification
How the classifier decides what to skip.
Manual upload
Drag PDFs in and bypass the classifier.
Forwarding addresses
Forward emails directly into Inbox Ledger.
Ready to try this?
Inbox Ledger turns your inbox into clean accounting data. The free tier includes 10 credits, refilled every 30 days.
Start free