Skip to content

Extract text with OCR

OCR (/ocr) turns scans and PDFs into editable text for classical Asian scripts.

  1. Open OCR from Tools.
  2. Upload a PDF, JPEG, PNG, or WEBP.
  3. Choose an engine / mode when offered — e.g. Auto, BDRC, or Surya via Dharmamitra MITRA OCR — and script options such as IAST or Wylie where relevant.
  4. Run OCR and copy or send the result into Translate or your notes.

When configured, an optional vision-LLM path can help on difficult images if the primary OCR path fails or returns thin output.

  • Prefer high-contrast, upright scans.
  • Crop to a single column when possible.
  • Check rare characters against the Dictionary or the printed source.