The Personal Knowledge Vault: Art Collection Use Case
summary
A deployed prototype turns 19th century ledgers and modern emails into linked graph documents, indexes them with a Neo4j-powered GraphRAG, and serves images from a MinIO-compatible image store so visitors scan QR codes in galleries to retrieve provenance, transcriptions and curator-written narratives. OCR and handwriting transcription run through asynchronous Celery workers, chunks are stored at ~500-token granularity, agents use Model Context Protocol and Cypher generation to answer queries and build custom tours; the system is live for institutional rollout next month.
AI-written summary · about
chapters (8)
- 0:00Introduction and background
- 3:03Art collection digitization
- 7:35Technology foundations and security
- 12:53Model context protocol explained
- 16:11Live demo of art retrieval
- 20:48Administrative interface and curation
- 24:15API layer and orchestration
- 27:05Q&A and future plans
related talks
23mThe Personal Knowledge Vault
40mContext That Dreams
42mBeyond Vectors: Evolving GenAI through Transformative Tools and Methods
32mBuilding an Advanced Knowledge Assistant
38mLeveraging the OPEA Platform and GraphRAG Architectures To Drive Cl...
45mSafely Capture Institutional Knowledge Using
40mTurn 10,994 Notes Into Memory
40mKnowledge graphs --enter- the Hype Cycle