AI2026
Scriptoria
Archive OCR for pre-reform print
Situation
Pre-reform texts didn't read with ordinary OCR.
Task
Put archive recognition on a conveyor.
Action
A recognition stream, a results feed, manual review of doubtful spots.
Result
The archive becomes searchable text, not pictures.
OCRPythonPipeline
Interfaces reconstructed and anonymized from client work where required.

