Vector and Search overview

Chromia combines PostgreSQL extensions, embedded vector stores, and on-chain access control to let you mix transactional and AI-centric workloads. This section introduces the primitives that power semantic search, recommendations, and hybrid filters.

Capabilities at a glance

• Native support for the pgvector extension with deterministic replication across providers.
• Full-text indexes that let you scope search queries to tenants or application chains.
• Deterministic execution for ranking functions so the same query returns the same ordered results across replicas.

Concepts to understand

Concept	Description
Embeddings	Store high-dimensional vectors per account, message, or object.
Hybrid search	Combine vector similarity with structured filters from the relational schema.
Access control	Use Rell to guard both writes and read access to sensitive embeddings.

Ready to build? Continue with the sample workloads.