When should I pick turbopuffer over LanceDB?

Pick turbopuffer if: You want a hosted index with nothing to operate.

Which do more makers use for vector database?

Among makers' products we track: LanceDB 3, turbopuffer 8.

LanceDB vs turbopuffer

Two sides of the vector database decision: embedded / local-first and managed vector store. When each fits, what it costs, who moves from one to the other, and what makers who chose it say.

LanceDBEmbedded / local-firstllms.txt

turbopufferManaged vector storellms.txt

Ask your AI about this, with this page as the source:ChatGPT ↗Claude ↗Perplexity ↗

Which fits you

Choose LanceDB if

An embedded vector database that stores vectors and data as files on local disk or object storage, with no server to run.

Use it whenYou want vectors in your app process or a serverless function, backed by S3-style storage.

Trade-offMany concurrent writers need care, and the managed cloud is a separate paid product.

Choose turbopuffer if

You want a hosted index with nothing to operate

Use it whenYou have many separate namespaces, such as one index per customer.

Trade-offNo free tier (paid plans have a monthly minimum), and it's cloud-only.

At a glance

	LanceDB	turbopuffer
Used by	3 makers' products · 39 open-source projects	8 makers' products · 7 open-source projects
Cost at default usagevectors stored 1 million vectors, queries 1 million queries, vectors written or updated 500k writes	—	$21/mo Launch
Downloads	1.5M/wk4.7× vs npm	1.1M/wk3.9× vs npm
Pricing	Free (open source); LanceDB Cloud billed by usage, Enterprise by contract.	Usage-based tiers with a monthly minimum; higher tiers add SSO, audit logs, and enterprise/BYOC deployments. · paid from $16/mo minimum usage
Free tier	Yes	No
Open source	Yes · self-hostable	No
Incidents, 90 daysfrom its status page	no public status feed	4 (4 major)

Loved and watch-outs

Themes that recur in makers' words and Hacker News comments, each linked to what it summarises.

LanceDBNothing that recurs in what we collected yet.

turbopuffer

Most loved

Storing vectors on object storage with a cache in front keeps large-scale search cheap. HN HN 2 HN 3 turbopuffer.com
Handles hundreds of millions to billions of documents, a common next step when pgvector runs out. turbopuffer.com turbopuffer.com 2 turbopuffer.com 3 HN
Operationally simple, with BM25, attribute filtering, recall tuning and consistency options built in. HN HN 2 HN 3 HN 4

Watch-outs

There is no offline local emulator, so development and CI must hit the hosted service. HN HN 2 HN 3
Its paid-only pricing floor is steep for small projects, which often stay on pgvector instead. HN HN 2

Who uses each

LanceDB3 makers' products

cognee Meet Dosu

CodeRabbit

turbopuffer8 makers' products

Mem0 Playerzero Pylon Readwise

fal Kepler Legora

Linear

What makers pair each with

With LanceDB

With turbopuffer

Transactional Email

Frontend Framework

Database

Neither quite right?

See the whole vector database decision →

pgvectorInside your databaseApps already on Postgres that want vector search in the same database, joined with normal tables.

ChromaEmbedded / local-firstPrototyping RAG on your laptop with a pip or npm install and no server to run.vs turbopuffer →

PineconeManaged vector storeA fully hosted index with nothing to operate, sized by usage.vs turbopuffer →

QdrantOpen-source vector storeHeavy metadata filtering alongside vector search, self-hosted from one binary or on its managed cloud.vs LanceDB →vs turbopuffer →

WeaviateOpen-source vector storeHybrid search that mixes keyword and vector results, with built-in modules that can create embeddings for you.