SimpleVecDB

SimpleVecDB

The dead-simple, local-first vector database.

1 follower

The Chroma experience in a single SQLite file. SimpleVecDB brings zero-infrastructure vector search to local RAG apps, offline agents, and indie hackers. Powered by usearch (HNSW) for sub-millisecond queries and 32x storage compression, it features hybrid BM25+vector search, full async support, and native LangChain/LlamaIndex integrations. Just pip install and go—no Docker, no cloud bills, no Redis. Your vectors live in one portable .db file that runs anywhere Python runs.
SimpleVecDB gallery image
Free
Launch Team / Built With
AppSignal
AppSignal
Get the APM insights you need without enterprise price tags.
Promoted

What do you think? …

malu
Maker
📌
Hey Product Hunt! 👋 I built SimpleVecDB because I kept hitting the same wall: every time I wanted to add vector search to a side project, I had to spin up Chroma, Pinecone, or Weaviate—which meant Docker, cloud accounts, or heavyweight dependencies. **The problem:** Vector databases shouldn't require infrastructure for local AI apps. **The solution:** SimpleVecDB is just a single SQLite file. No servers. No containers. No monthly bills. What makes it different: • ⚡ ~2ms query latency on consumer hardware • 📦 32x storage reduction with BIT quantization • 🔍 Hybrid search (BM25 + vector) out of the box • 🔌 Drop-in LangChain & LlamaIndex adapters • 🔒 Your data stays local—perfect for sensitive workloads It's powered by `sqlite-vec`, which means your vectors live in the same portable .db file that powers iOS apps, browsers, and embedded systems worldwide. **Who's it for?** → Indie hackers building local RAG apps → Teams prototyping without cloud overhead → Privacy-focused projects that can't send data to external APIs The entire project is MIT licensed and open source. I'd love your feedback on what features would make this more useful for your workflows!