PGSync adds semantic search without data leaving your infrastructure
8 hours ago
- PGSync Pro provides real-time semantic and hybrid search for PostgreSQL, syncing vectors via CDC.
- It only re-embeds content when the actual meaning changes, saving on embedding costs.
- Combines vector kNN and keyword BM25 using reciprocal rank fusion for hybrid search.
- Supports local sentence-transformers embeddings or optional cloud providers like OpenAI.
- Includes an MCP server for agent integration (e.g., Claude Desktop) and a preflight validation tool.
- PGSync and django-pgsync remain free MIT; Pro adds the semantic embedding layer.