How-to
Translating a PDF without uploading it
· 1 min read
Search “translate a PDF” and almost every result asks you to upload the file. For a holiday menu that is fine. For a contract, a medical record, a legal filing, or an unpublished manuscript, the upload is the problem — the document is exactly the thing you cannot hand to a stranger’s server.
The file never leaves the browser
On the PDF page, the file is opened and parsed by your own browser. Its text is translated by the on-device engine, page by page, and the translated layout is rebuilt to match the original. The bytes of the PDF are never sent anywhere — there is no code path that could upload them.
Scanned PDFs, too
A scan is a picture, not text, so it needs OCR first. That recognition also runs locally, in the browser. The one thing that touches the network is a small, one-time download of the recognizer’s language data — never your document. The page tells you the first time it needs it.
Steps
- Install a language pack once on the Translate page (about 1.5 GB, then offline forever).
- Open the PDF page and drop your file in. It stays on your device.
- Watch it translate page by page; edit any block in place if you want to.
- Download the translated PDF, or copy the plain text — whichever you need.
Because it is all local, it works on a plane, behind a corporate firewall, or on a network you do not trust. The document that was too sensitive to upload is the one this was built for. If you are comparing local PDF tools, the translate-a-PDF-locally page goes deeper.