OCR PDF
Character recognition for scans — Turkish and English — running as WebAssembly in your browser. The scan never leaves your device.
Choose a scanned PDF
or drop files here
▚ files open locally — nothing is uploaded
The OCR engine runs as WebAssembly in your browser — the scan is never uploaded, in any step.
- 01Open a scanned PDF and pick the document language (Turkish + English is the safe default).
- 02Click Recognize. The OCR engine loads once (~10 MB, then cached) and reads each page on your device.
- 03Download the searchable PDF — it looks identical but the text can be selected, copied and searched — or grab the plain .txt.
Scanned documents are where privacy and utility collide hardest: the files are images, so every OCR service says "upload it and we'll read it for you." For medical records, court files and old contracts, that trade is exactly wrong.
ZorPDF runs Tesseract — the industry-standard OCR engine — compiled to WebAssembly, in your browser. Your scan is rendered, read and rebuilt as a searchable PDF on your own hardware. The result: a document that looks identical, but suddenly answers to Ctrl+F, in Turkish and English alike.
How is this different from other OCR sites?+
Every other OCR service we know of uploads your scan to a server farm. ZorPDF runs the recognition engine as WebAssembly inside your browser — the scan, and everything it says, stays on your device. You can watch the Network tab while it works.
What does 'searchable PDF' mean?+
The output looks exactly like your scan, but an invisible text layer sits over each word — so Ctrl+F finds things, text can be selected and copied, and the file becomes indexable.
Which languages are supported?+
Turkish and English, separately or together. Turkish characters — ğ, ş, ı, İ — are recognized and embedded correctly.
How accurate is it?+
Very good on clean scans at normal sizes; expect occasional misses on low-resolution photos, handwriting or unusual fonts. The engine is Tesseract, the same open-source OCR used industry-wide.
Why is the first run slower?+
The engine (~10 MB of WebAssembly and language data) downloads on first use — from our server, not Google's or anyone else's — and is cached after that. The recognition itself uses your device's CPU.