Preserving the written heritage of Sambas through Artificial Intelligence.
SambasJawi AI combines deep-learning OCR, expert validation, transliteration, archival metadata, and digital preservation to transform historic Malay Jawi manuscripts into searchable and trusted cultural knowledge.
From physical manuscript to trusted digital heritage
Every manuscript follows a traceable workflow that combines preservation science, artificial intelligence, and expert scholarly validation.
Digitize
Capture manuscript pages in preservation-grade quality.
Enhance
Prepare images through quality control and preprocessing.
Recognize
Run the SambasJawi deep-learning OCR engine.
Verify
Jawi experts review and certify AI transcription.
Transliterate
Convert verified Jawi text into Latin representation.
Preserve
Store master files, metadata, provenance, and checksums.
Publish
Release approved heritage collections to the public repository.
AI accelerates the work. Experts protect the meaning.
SambasJawi AI is designed around human-in-the-loop validation. OCR predictions are never treated as historical truth until a qualified validator reviews them. Corrections are preserved as provenance-rich ground truth for future model improvement.
Cultural Preservation
Protect fragile manuscripts through high-quality digitization, archival metadata, integrity checks, and controlled publication.
Research Infrastructure
Support scholars with validated transcription, transliteration, annotations, model metrics, datasets, and reproducible provenance.
Public Knowledge
Make verified heritage collections searchable and accessible for education, cultural learning, and future generations.