Tibetan Woodblock Digitization

We are digitizing Tibetan block-print manuscripts (pecha): taking scans of the original woodblock folios and converting them into searchable digital Tibetan text — together with a transliteration, a simple pronunciation guide, and a draft English translation.

Everything runs locally: OCR uses the BDRC woodblock models, pronunciation and Wylie are derived with open tools (bophono, pyewts), and the draft translation comes from a small translation model trained on Buddhist texts (MLotsawa). The digital text is then proofread against the original print, folio by folio.

How a text moves through

1. Scan — high-resolution images of the woodblock folios
2. OCR — BDRC Woodblock model reads the print into Unicode Tibetan
3. Clean — normalization and mechanical fixes
4. Layers — Wylie transliteration, pronunciation, draft English
5. Proofread — checked line-by-line against the scan by a qualified reader

Texts