pgvector
https://github.com/pgvector/pgvector
Stores embeddings alongside relational + full-text data. HNSW + IVFFlat ANN indexes. Used as agent conversation memory via LangChain + MCP. Foundation of Supabase AI and many self-hosted RAG stacks.
At a glance
- Type
- PostgreSQL vector extension
- Tier
- T1
- Section
- Vector-database infrastructure
- Created
- 2021-04
- Latest release
- no releases
- License
- Custom
- GitHub
- 21.1k★ +123/mo C
- Pricing
- Free (PostgreSQL License open source)
- Funding
- No external funding — open-source project by independent developer
Taxonomy
- storage
- vector
- retrieval
- similarity
- persistence
- long-term
- update
- overwrite
- unit
- chunk
- governance
- inspectable
- conflict
- overwrite
When to use
Optimised for: low-latency similarity search + scale
Anti-fit: not for relational / graph-heavy queries (vector-first by design)
Pros & cons
Pros
Postgres extension — keeps vector data alongside transactional data with full ACID; minimal ops for shops already running Postgres.
Cons
Performance plateau at very large vector counts; limited filter pushdown vs purpose-built engines.
Claims & capabilities
PostgreSQL license. Most-deployed open-source vector store.
Technical surface
- API surface
- SQL via Postgres protocol
- Backend storage
- Postgres
- Deployment
- Self-hosted only (PostgreSQL extension; managed via Supabase/AWS RDS/etc.)
- Embedding model
- BYO
- Multi-tenancy
- Logical isolation via Postgres schemas/databases; managed providers (Supabase, Neon, RDS) add tenant-level controls
- MCP
- via Postgres MCP servers (community)
- A2A
- not supported
- OpenTelemetry
- via Postgres OTel
Compare pgvector with…
Similar systems
Other vector-database infrastructure in the catalog, ranked by inbound references.
- Qdrant T1
Distribution-Based Score Fusion + RRF. Sparse vectors native; filtering via ANN graph modification.
- Pinecone T1
Managed vector DB. Cascading sparse + dense + rerank pipeline; pinecone-rerank-v0 .
- Chroma T1
Limited native hybrid (users build RRF custom). Fast Rust core (v2.5).
- LanceDB T1
Embedded vector DB (Arrow columnar). RRF reranker. Petabyte-scale on disk.
- MongoDB Atlas Vector Search T1
Agent memory store for both short-term (document) and long-term (vector). LangGraph checkpointer for stateful agents. Vector search extended to Community Edition (Sept 2025).
- Milvus T1
Multi-vector columns (10 simultaneous). Native hybrid search (v2.5). CAGRA + Vamana GPU/CPU (v2.6).
Related systems
References (3)
- LangChain (framework) integrates with — Used as agent conversation memory via LangChain + MCP.
- LangChain (framework) depends on at runtime — conversation memory via LangChain + MCP. Foundation of Supabase AI and many self-hosted RAG s
- PostgreSQL pgvector same team as — Same GitHub organisation `pgvector` publishes both repos.
Referenced by (5)
- Agno (Phidata) Memory builds on — Postgres + pgvector default — operational simplicity for shops that already run Postgres.
- Apache AGE integrates with — openCypher over Postgres. Pairs with pgvector for graph + vector hybrid retrieval.
- Letta / MemGPT depends on at runtime — backend-storage cell: Postgres + pgvector
- Mem0 Security / OpenMemory builds on — OpenMemory is the local self-hosted variant (Docker + FastAPI + Postgres + Qdrant)
- Memobase depends on at runtime — backend-storage cell: Postgres + pgvector