Gateway to written heritage
Upload a scan and within moments you have a transcription, modern language translation, historical context, and glossary — whether the original is in Latin, Old Czech, or German.
0
Pages uploaded
0
Documents processed
0
Volumes
0
Languages
Three steps
01
Upload a scan or photograph of a historical text. We support JPEG, PNG, TIFF, WebP and PDF. You can also paste a URL from digital libraries.
02
A multimodal language model with transformer architecture performs palaeographic raster analysis, historical text tokenisation, and cross-lingual translation with contextual annotation.
03
Using the integrated chat, you can ask the model anything about the document and have it correct the transcription or translation.
For power users
With the ais tool you can upload scans, run OCR locally using your own Claude subscription, and edit results directly in the terminal — perfect for integration with AI assistants and everyday editors.
Processing runs on your machine via Claude CLI — you pay Anthropic directly, no middlemen.
Download transcriptions as markdown files, edit in any editor, and push changes back to the server.
Claude Code, Cursor, Copilot — let AI fix transcriptions, translations, and context. Or edit manually, your call.
npm install -g @ai-sedlacek/cliSpecialisation
Bastarda, cursive, 14th–16th c.
Fraktur, Schwabacher, chancery script
Medieval manuscripts and prints
Multilingual documents, glosses
About
The project is named after August Sedláček (1843–1926), a Czech historian and archivist who dedicated his life to the study of historical sources. Our goal is to make medieval texts accessible to modern readers.
Try for free