HeritageLM

LLM-powered knowledge extraction that unlocks large collections of historical and cultural documents.

Category
LLMs for cultural heritage
Sector
Culture · Archives
Status
Concept
Readiness
TRL 2 of 9

HeritageLM is an LLM-powered knowledge extraction platform designed to unlock the value of large collections of historical and cultural documents. It digitises, interprets and semantically organises newspapers, museum archives, manuscripts, books, historical records and administrative documents using OCR, NLP and LLMs.

It generates structured metadata, searchable digital archives, multilingual summaries, semantic search and knowledge graphs, accelerating digital preservation, archival research and public access to cultural heritage.

What HeritageLM does.

01

OCR & digitisation

From scanned pages to searchable text.

02

Structured metadata

People, places, dates and topics extracted automatically.

03

Semantic search

Ask questions across the whole archive, in several languages.

04

Knowledge graphs

See how documents, people and events connect.

Research deeply. Build responsibly. Make it real.

Want to see HeritageLM in action?