Pinecone

A managed vector database that handles the infrastructure so you don't have to. You get sub-100ms p95 latency, auto-scaling to billions of vectors, and hybrid search with dense and sparse vectors out of the box. The serverless option is genuinely convenient for production RAG apps where you want predictable performance without tuning pods or managing clusters. Integrates cleanly with LangChain and LlamaIndex. Free tier gives you 100K vectors to prototype with, which is enough to know if it fits your use case. The tradeoff is cost and vendor lock-in compared to self-hosting Chroma or Weaviate, but that's the deal with any managed service. Metadata filtering works well for multi-tenant scenarios using namespaces.

Install to Claude Code

npx -y skills add orchestra-research/ai-research-skills --skill pinecone --agent claude-code