QQuantumSteel Design

Your browser. Your workspace.

Local Scanned Drawing OCR

Recognize English text on selected scanned PDF pages locally, compare the text with the page image, and export a corrected transcription without uploading drawings.

OCR turns pixels into a draft transcription. Check every dimension, fraction, symbol, and mark against the source image before using it. For existing embedded PDF text, use Drawing Set Search first.

Choose pages for local recognition

No PDF selected.

The first run downloads about 15 MB of software assets from this website. Your drawing bytes and recognized text stay on this device. Large drawing sheets are rendered at a maximum of 6 million pixels; tiny text may need a cropped PDF export for better recognition.

Review recognized text

Recognize a page to search its text.

A careful OCR workflow

  1. Select only the pages needed for the current question. Start with a clear notes page to judge the scan quality.
  2. Choose scattered labels for drawings or one text block for a tightly cropped notes section. Recognition uses the English model; it does not translate text.
  3. Compare the generated transcription with the image beside it. Edit the text when a mark or symbol is wrong. Confidence is an engine estimate, not a quality certificate.
  4. Export the edited text with the original OCR output for traceability. The source PDF is left unchanged, and this tool does not add a searchable layer to it.

OCR can confuse 0 and O, 1 and I, fractional dimensions, diameter symbols, weld symbols, and rotated text. A missing word can change the meaning of a note. Never use an OCR result alone to choose a connection, approve a release, or fabricate a member.

Can I cancel a long recognition run?

Yes. Cancel stops the OCR task, including its initialization worker, and keeps completed pages. Clear also releases the PDF and removes completed text.

Is this an online document service?

No. PDF rendering and Tesseract recognition run in your browser using software hosted on this website. No drawing, password, transcription, filename, or query is sent to an OCR service. Closing this page discards the transcription unless you exported it.

OCR engine: Tesseract.js. PDF rendering uses the same locally hosted PDF.js engine as the other drawing tools.