OCR a payslip without sending it to the cloud

Scanning a payslip to make it searchable is a 30-second task — except every popular online OCR sends the document to a server first. A payslip contains the employee's name, address, bank account number, social security number, and net salary. Uploading that to a free SaaS is a poor deal. iSharePDF runs Tesseract OCR compiled to WebAssembly inside the browser. The payslip is processed on your CPU, not on someone else's.

→ Open the OCR tool — no upload

Why iSharePDF for OCR on confidential documents

Payslips, bank statements, medical reports, tax returns — they share a profile: scanned, often image-based, full of personal identifiers, and frequently OCR'd as part of bookkeeping or onboarding. For each of these, the upload-based OCR tool is the wrong tool:

No third-party processor. The OCR runs inside your browser tab; there is no server in the pipeline, so there is no DPA to amend.
No retention. Most online OCR services state "files deleted after 1 hour" — but they were uploaded for that hour. Local OCR means zero retention window because there is no upload.
No usage-based pricing. Tesseract WASM is free; iSharePDF gives you 5 OCR runs per day for free, unlimited on Pro at €4.99/mo. No per-page billing tied to your document's content.

How it works in 4 steps

Open isharepdf.com/app/tools/ocr — no signup.
Drop your scanned payslip PDF (or JPG). The file loads instantly because no upload happens.
Pick the language pack (English, French, or 130+ in Pro). First-time downloads the language model (~10 MB) from a public CDN.
Click "Run OCR". Tesseract processes each page on your CPU. Download the searchable PDF when done.

Verify it yourself with DevTools

Open the OCR tool with DevTools' Network tab open. You will see one network request: the initial fetch of the Tesseract language model (a public file, not specific to you). Then your payslip is processed — zero further outbound requests carrying its content. Compare with online OCR services like onlineocr.net or Adobe Acrobat OCR: those will POST your payslip to their server.

OCR accuracy on payslip-style scans

Tesseract's accuracy depends on scan quality:

200 DPI or above: 95%+ character accuracy on standard payslips. Numerical fields (net salary, IBAN) are typically clean.
120-200 DPI mobile photo: 80-90%. Manual review of IBAN digits recommended.
Below 120 DPI: noisy. Re-scan at higher resolution if possible.

For French payslips with non-ASCII characters (é, à, ç) make sure to pick "Français" as the language pack. Mixed-language documents can stack multiple packs in Pro.

FAQ

Will my payslip be readable by anyone else after OCR?

No — the OCR runs on your machine. The searchable PDF output only exists on YOUR disk, the same way as if you had run OCR in Acrobat Pro locally. iSharePDF never sees the file.

Why does Tesseract download a language pack — is that a leak?

The language pack is a generic model file shared by all users of Tesseract worldwide. Your browser fetches it from a public CDN. Your specific document is not part of that request. After the first OCR you run, the pack is cached locally.

What about bank statements and tax returns?

Same architecture, same answer. Anything you'd OCR offline with Acrobat Pro, iSharePDF can do in the browser, with the additional benefit that there is no software install.

OCR a payslip — without sending it anywhere

One click. Tesseract runs on your laptop. Zero data leaves your machine.

Open the OCR tool →