Local Processing - No External AI Models Involved

Financial documents contain sensitive information - account numbers, transaction amounts, vendor details. Your document contents are read and processed locally in your browser: OCR, parsing and matching all run on your machine, with no external AI models involved, and the contents of your invoices, receipts and statements are not uploaded. Two analytics services do run on every visit: Google Analytics (measurement id G-XM5TP481ZK) records page views and anonymous product events such as a file being uploaded, a statement being parsed or an export completing, and Microsoft Clarity (project id vp2sfdbk62) records session replays, which are recordings of how the page is used. Analytics events have filenames, vendor names and document contents stripped out before they are sent, and the replays are captured with every text node on the page masked, so your filenames, vendor names and amounts are not readable in them. Separately, if matching goes wrong you can choose to send your files to the developer so the problem can be reproduced - that is an opt-in tick box in the export dialog, it is off by default, and nothing is sent unless you tick it.

What We Collect

• Your documents: nothing, by default. Invoices, receipts and bank statements are read in your browser. The files themselves, and the text parsed out of them, are not sent anywhere unless you opt in below.
• Anonymous usage data: on every visit. Google Analytics (G-XM5TP481ZK) collects page views plus anonymous product events - a file was uploaded, a statement was parsed, an export completed - with filenames, vendor names and document contents stripped out first. Microsoft Clarity (vp2sfdbk62) records session replays of how the page is used, with all page text masked so no filename, vendor name or amount is readable in the recording.
• Your files, only if you ask us to. If matching goes wrong, the export dialog offers a tick box that sends your files to the developer so the failure can be reproduced and fixed. It is opt-in and off by default.
• What you type into the contact form: stored and emailed. The name, email address, subject and message you submit on Contact are written to a database on igori.eu and emailed to the site owner so they can be read and replied to. Only a one-way hash of your IP address is kept, for rate limiting - never the address itself. Nothing else on the site sends anything you type to a server.

How It Works

PDF text extraction uses PDF.js, Mozilla's open-source PDF renderer that runs entirely in JavaScript. For scanned images and photo receipts, the app uses Tesseract.js, a WebAssembly port of the Tesseract OCR engine.

The matching algorithm compares extracted receipt data against your bank transactions using amount, date, and vendor name - all computed locally. No external API calls or AI models are used to process your files.

No Accounts, No Sign-ups

There is no login, no registration, and no user accounts. You open the page, drop your files, and download the result. Your session state is saved in your browser's local storage so you can resume later. The only personal data stored on the server is what you choose to send through the contact form.

ZIP Export

When you export your matched transactions, the ZIP file is assembled in-browser using JSZip and saved directly to your device.