Ledger
Financial PDFs in.
Clean spreadsheets out.
Ledger reads bank, credit card, 401k, brokerage, check-register, and general-ledger PDFs - text-based or scanned - and transcribes every line item into a ready-to-use Excel workbook. Exactly as printed. No re-keying, no copy-paste, no dropped rows.
Already a customer? Log in to Ledger
Reads the documents you actually get
- Bank statements
- Credit card statements
- 401k statements
- Brokerage statements
- Check registers
- General ledgers
- Scanned documents (OCR)
Built for forensic review, not just convenience
When statements become evidence or workpapers, "close enough" is not good enough. Ledger is designed to capture everything and change nothing - it would rather flag an ambiguous row than silently drop it.
Every line item
The extraction engine is instructed to capture every row on every page - multi-page statements, continuation pages, and all. Completeness first.
Exactly as printed
Values are transcribed as they appear - no rounding, no sign flipping, no "helpful" normalization. Your spreadsheet says what the document says.
Scans welcome
Image-only PDFs are read page by page with built-in OCR - no separate OCR tool, no pre-processing step. Drop the scan in like any other file.
One workbook per file
Each PDF becomes its own .xlsx with a Transactions sheet and a Summary sheet - institution, masked account number, statement period, balances.
Built for batches
Drag in a folder's worth of statements at once, or push up to 100 files per batch through the key-authenticated API and its PowerShell client.
Private by design
Single sign-on in front, AI extraction on Claude via Amazon Bedrock inside a private AWS environment - documents are never sent to a third-party SaaS.
Three steps. Minutes, not afternoons.
1. Upload
Drop in one PDF or a hundred. Optionally tag the batch with a document type - or let Ledger auto-detect what each file is.
2. Extract
AI reads every page - including scans - and returns structured rows with the right columns for each document type, plus account metadata.
3. Download
Preview the rows in the browser, then download each workbook, grab them all at once, or send a secure, expiring link by email.
Workbooks that name and file themselves
Every extraction produces a two-sheet Excel workbook: the full transaction detail, plus a summary of what Ledger found - institution, account (last four only), statement period, and opening and closing balances. Files are named from the document itself, so a hundred downloads stay organized.
- Transactions sheet: every extracted row, headers bolded, columns sized
- Summary sheet: document type, row count, and account metadata
- Smart names like
firstbank_x4821_2025-11.xlsx