TerminusDB 11.1.0 adds the VectorLink semantic indexer for AI-powered similarity search, and a concat command for merging storage layers. It was released on 10 July 2023.
- Semantic indexer — VectorLink indexes documents for similarity search.
concatcommand — merge base layers into one.- Lower memory usage — leaner server, no Google Analytics in the dashboard.
Semantic indexer
What it does
Support for the TerminusDB semantic indexer — VectorLink — a sidecar that stores, indexes and retrieves embedding vectors by similarity. Documents can be rendered to text, embedded, and searched semantically on top of the knowledge graph.
Why it matters
This VectorLink 1.0 brings Vector Search to TerminusDB and enables every structured document to get a semantic vector embedding automatically (GraphQL and Handlebars to render from a template), and at the same time having the document as a structural part of the knowledge graph too. Full git-for-data including for Vector Search is a powerful capability.
The roadmap version of VectorLink 2.0 will offer even more advanced features and capabilities. Check it out when released!
Read more: Set up VectorLink · Use VectorLink · OpenAI & Handlebars configuration
concat command
A new concat command merges base layers into a single layer — a manual, more aggressive variant of the automatic squashes the optimizer performs.
Lower memory usage
General memory reductions across the server, and the dashboard no longer ships Google Analytics.
Fixes
- Fixed GraphQL ordering of stringy numbers; added type information to decimals
- Corrected diffs using out-of-band values; prevented reserved words in GraphQL names
- Fixed loading of triples
Upgrade
docker pull terminusdb/terminusdb-server:v11.1.0No data migration is needed. Docker install guide · Full changelog on GitHub
Previous: TerminusDB 11.0.6: Automatic Schema Migrations · Next: TerminusDB 11.1.1: Multiple Inheritance for Properties · TerminusDB 11.1 changelog · Release index