Best forA fully hosted index with nothing to operate, sized by usage.
Trade-offClosed source and cloud-only, so leaving means re-indexing somewhere else.
Used by 55 makers' products for this
pgvector alternatives
Not a ranking: each option is described by the situation it fits, with its trade-off, whether it's open source or free to start, whether AI coding agents can work with it, and how many makers' products use it for this job.
7 open source or self-hostable9 with a free tier9 ready for AI agents
For vector database
The whole decision →Inside the database you already run, or a separate vector store. pgvector keeps vectors next to your rows, with one backup and one query language, but tuning large indexes is on you. A dedicated vector store handles scale, filtering and hybrid search for you, at the cost of another service to sync data into.
pgvector (Inside your database): Apps already on Postgres that want vector search in the same database, joined with normal tables. Trade-off: Index tuning and memory sizing are your job, and very large indexes can crowd out the rest of the database.
Best forHeavy metadata filtering alongside vector search, self-hosted from one binary or on its managed cloud.
Trade-offOne more service to deploy and keep in sync with your main database.
Used by 18 makers' products for this
Best forHybrid search that mixes keyword and vector results, with built-in modules that can create embeddings for you.
Trade-offMore concepts and configuration to learn than simpler stores; some advanced features need a license key.
Used by 16 makers' products for this
Best forPrototyping RAG on your laptop with a pip or npm install and no server to run.
Trade-offFor production you either run its server yourself or move to Chroma Cloud.
Used by 13 makers' products for this
From $16/mo minimum usagellms.txt
Best forVery large or many-tenant indexes where storing everything on object storage keeps cost down.
Trade-offNo free tier (paid plans have a monthly minimum), and it's cloud-only.
Used by 8 makers' products for this
Best forAn embedded vector database that stores vectors and data as files on local disk or object storage, with no server to run.
Trade-offMany concurrent writers need care, and the managed cloud is a separate paid product.
Used by 3 makers' products for this
Best forApps already on MongoDB Atlas that want vector search on the same documents, with no second store to sync.
Trade-offVector search is most mature on Atlas; on self-hosted MongoDB it arrived later, so check your version and setup.
No maker product tracked for this yet
Best forA serverless vector index billed per request, alongside Upstash Redis and QStash, reached over HTTP from edge functions.
Trade-offClosed source and hosted only; per-request pricing gets expensive at very high query volume.
No maker product tracked for this yet
Best forVery large collections — hundreds of millions of vectors and up — on a distributed open-source database, or managed as Zilliz Cloud.
Trade-offThe distributed deployment has many moving parts; for small apps it's more system than you need.
No maker product tracked for this yet
Open sourceFree tier
Best forA fast in-memory vector index library for research, batch jobs and apps that load their index at startup.
Trade-offA library, not a database — no metadata filtering service, persistence or updates beyond what you build.
No maker product tracked for this yet