Vector field and search capabilities using pgvector.
Use PostgreSQL as your vector store — no external services needed
Vector similarity search for PostgreSQL
pgvector is a PostgreSQL extension that adds support for vector similarity search. This module implements vector storage directly in your Odoo database, eliminating the need for external vector database services. Your embeddings stay alongside your data.
Vector capabilities built into PostgreSQL
Vector field type supporting variable dimensions to match your embedding model requirements.
Store and retrieve embeddings for knowledge chunks directly in PostgreSQL.
Manage vector indexes for optimized search performance on large datasets.
Perform fast similarity searches leveraging pgvector's optimized algorithms.
Benefits of using PostgreSQL for vectors
Use your existing PostgreSQL database. No need to deploy and maintain separate vector databases.
Embeddings stay in your database alongside your business data. Better for security and compliance.
Vector operations participate in PostgreSQL transactions. ACID guarantees for your AI data.
Requirements and dependencies
llm, llm_knowledge, llm_store
pgvector, numpy
pgvector (auto-installed)
LGPL-3