LiveMessy exports in. Clean books out.
Ledger Lens turns bank-statement exports and receipts into categorised, accountant-ready books — every transaction tagged, deduped, and explained. Rules do the obvious work offline; AI only fills the gaps.
Why
Startup books are a pile of CSV exports, PDF receipts, and half-remembered merchants. Categorising it by hand is slow — and black-box tools just guess, with no way to see why.
How it works
Drop your exports
Upload bank statements (CSV / XLSX) and receipts (PDF / image). Everything is parsed locally in your browser — nothing is uploaded.
Rules categorise, you review
A deterministic engine tags known merchants offline and shows why — rule, AI, or manual. Every category is editable, and your corrections are remembered.
Export accountant-ready books
Get a categorised ledger and a category summary as clean CSVs, plus a reusable rules.json of your corrections for next time.
Features
Rules-first — free and offline
Known merchants are categorised by a local rules engine. No key, no cost, no data leaving your machine.
AI only for the remainder
Bring your own key — Anthropic, OpenAI, Google, or Groq. The AI classifies only what the rules couldn't, and never touches amounts.
Audit-grade, never a black box
Every row shows its basis: which rule fired, the AI's confidence, or your manual override.
Receipts reconciled
PDF and image receipts are matched to their statement line by amount and date, so they enrich a row instead of duplicating it.
Dedupe + multi-currency
Overlapping exports are de-duplicated automatically. Read totals in ₹ or $.
Corrections that stick
Fix a category once and it's remembered across rows and re-imports — exportable as rules.json.
Who it's for
Built for early-stage founders and operators who close their own books before handing clean files to an accountant.
Open & honest
FAQ