Powering next gen
 AI apps with Postgres
Build and scale transformative LLM applications with vector indexes and similarity search in Postgres
Get StartedScale your AI apps to millions of users with Neon


Speed up your queries with HNSW
Query execution time (ms) at 99% recall
HNSW indexes bring 20x the speed for 99% accuracy to graph-based approximate nearest neighbor search in your Postgres database.
 Uses HNSW
Uses HNSWSimple to use,
 scales automatically
Store vector embeddings and perform similarity search
CREATE EXTENSION vector;
CREATE TABLE items (id BIGSERIAL PRIMARY KEY, embedding VECTOR(3));
INSERT INTO items (embedding) VALUES ('[1,2,3]'), ('[4,5,6]');
SELECT * FROM items ORDER BY embedding <-> '[3,1,2]' LIMIT 1;Store embeddings and perform vector similarity search in Postgres with pgvector.Learn moreabout pgvector
Vector search with Neon
Use the power of HNSW indexes to unlock new levels of efficiency in high-dimensional vector similarity search in Postgres
- Reliable & actively maintained- The pgvector extension is open-source and actively maintained 
- Amazing scalability- Grow your vector stores without impacting search performance 
- Blazingly fast search- Use HNSW indexes for fast and scalable vector similarity search in Postgres 
- Highly compatible- Use Neon with pgvector in your Postgres and LangChain projects 
Start building
 AI apps on Neon
Check out the following example LLM and AI apps and start building in minutes
- Chatbot- Give LLM-based chatbots long-term memory and provide relevant context from your data. View example
- Semantic search- Build next-level search experiences for your users where you understand the true meaning of their queries. View example
- Image similarity search- From the Neon Community: Image similarity search with Vertex AI, Neon, and pgvector. View example
Build your next AI app now with Neon
Build your AI app on our Free Tier. Upgrade to a paid plan when you're ready to scale.

