Back

Searchable scanned PDF

Document tools

Loading

Loading tool

The tool is loaded only when you open it.

All processing for this tool happens in your browser. Your input is not sent to a server.

About this tool

Recognize a scanned PDF locally with the selected model. All pages finish text and visual checks before the new PDF is available. Your PDF is never uploaded.

Common uses

  • Add a verified invisible English, Chinese or Japanese OCR layer to a local scanned PDF.
  • Review recognized text and search or copy it in a compatible PDF reader.

How to use it

  1. 1.Recognition model — Choose a scanned PDF
  2. 2.Recognize and verify PDF
  3. 3.All 1–10 pages passed text and visual verification. Review the recognized text and the exported PDF. Save searchable PDF

Limits and notes

  • One PDF ≤20 MiB, up to10 pages; output ≤30 MiB. Raster ≤2000px per edge/4 million pixels. Embedded images ≤16 million pixels each/64 million total; ≤20000 PDF objects. Read/render/export each30s; model60s and OCR120s per page; processing total600s. These limits do not cap all native/PDF/WASM memory. Decoded content streams ≤16 MiB total; only raw/Flate content. Inline images and input Type3 fonts are rejected. JPEG must be8-bit grayscale/RGB baseline/progressive with matching declared/frame dimensions; image filter chains are unsupported. Contents arrays must be flat with≤20000 streams.
  • For clear printed scans with no existing text. English, Simplified Chinese or Japanese model; no remote fallback. Review every character: OCR and word positions are estimates; handwriting, vertical text and complex tables are outside the tested scope. Search-layer whitespace becomes single spaces; preview escapes hidden controls, copying retains original OCR text. The new PDF keeps page appearance but loses document metadata, bookmarks, tags and signatures. Forms, nonempty annotations, encryption, optional layers and unsupported image codecs are rejected. Not redaction or an archival/accuracy guarantee. Fixed engine/model/font assets load from this site; no input persistence or model IndexedDB cache. New file/model/locale, clear, cancel or leaving discards results.

Frequently asked questions

Does my PDF leave this device?

Recognize a scanned PDF locally with the selected model. All pages finish text and visual checks before the new PDF is available. Your PDF is never uploaded.

What if verification fails?

Text or appearance could not be verified. No PDF was exported.

Related tools