What it is for
Palaeography is the study of historical handwriting. Archetype lets a researcher record what they see in a manuscript hand — the shape of an a, the way a stroke was made — and turn those observations into evidence they can search, compare and cite, anchored to the image of the page. It is open software: any project can run it over its own manuscripts.
The corpus it is proving itself on is Models of Authority, Scottish charters from 1100 to 1250: 710 charters, some 3,000 page images and 12,700 annotated letter-forms, live in public beta.
Why a rewrite
Archetype grew out of DigiPal, a framework built at King’s College London more than a decade ago. The research it made possible was real; the software had reached the end of what it could be maintained into. Version 3, in development since 2024, keeps the ideas and starts the code again: a Django and Django REST Framework API with Celery for the slow work, PostgreSQL, Meilisearch for search, a IIIF image server, and a TypeScript frontend, all run with Docker so an institution can host its own.
It speaks the standards the field already uses: IIIF for images, TEI for transcriptions and W3C Web Annotations for what researchers mark on a page.
My part
I lead the engineering and have written most of the backend and the frontend, along with the infrastructure that deploys the beta. The day-to-day is what you would expect of a small open-source team with research deadlines: shaping the API and the search contract with the people who use them, keeping the test suite and the CI honest, and leaving documentation behind for the next contributor. Version 3 is supported by the University of Glasgow’s OG(H)AM project and the École Pratique des Hautes Études.