Pinecone
Managed, serverless vector database for semantic search and RAG, with a free starter tier.
- #vector-database
- #rag
- #embeddings
- #search
Pinecone is a fully managed vector database. You upsert embeddings with metadata and query by similarity, and it handles indexing and scaling.
Why it’s useful
- Serverless: storage and compute scale separately, and you pay for what you use.
- Size it first: the number of vectors you store equals your chunk count. The RAG Chunk & Token Calculator estimates it from your corpus size, chunk size and overlap.
- Hybrid (sparse + dense) search, metadata filtering and namespaces.