Skip to content
TrustList
Products · category

Top vector databases

20 vector database software ranked by rating, review volume and listing quality.

Want your product here?

What to look for in vector databases

The best choice depends on your workload, your team and your budget, not only on where an engine sits in this list. Before you shortlist the top-ranked vector databases, check:

  • Filtering on metadata together with similarity search, which most real applications need
  • Index types and the trade-off between recall, speed and memory they let you tune
  • Hybrid search: combining keyword and vector results
  • Scale: how many vectors and how many queries a second the deployment you can afford will hold
  • Self-hosted and managed options, and the licence of each
  • Whether the database you already run (PostgreSQL, Elasticsearch, Redis, MongoDB) offers vector search good enough to avoid a new system

Vector search is also offered inside general-purpose engines listed in other rankings: PostgreSQL through the pgvector extension, Elasticsearch and OpenSearch, Redis, and MongoDB Atlas. For many teams one of those avoids running a separate system.

How to use this ranking

Use it to build a shortlist. Read each engine’s profile for its licence, deployment options and latest release, then test the two or three that fit on your own data and queries before you commit.

Filter by fitbudget, engagement, size and more

Only businesses that have told TrustList match a filter. Filters never change the order or a score.

  1. P

    Pinecone

    DataBase

    Pinecone is a proprietary, fully managed vector database service for teams that want hosted similarity search without running an engine themselves.

    Write a review
  2. M

    Milvus

    DataBase

    Milvus is an open-source vector database for storing and searching embeddings, deployable on your own infrastructure or through Zilliz Cloud.

    Write a review
  3. ZC

    Zilliz Cloud

    DataBase

    Zilliz Cloud is a proprietary, fully managed vector database service powered by Milvus, for teams that want Milvus without operating it.

    Write a review
  4. Q

    Qdrant

    DataBase

    Qdrant is an open-source vector search engine for similarity search with payload filtering, available self-hosted or as a managed cloud service.

    Write a review
  5. W

    Weaviate

    DataBase

    Weaviate is a vector database with hybrid search, available self-hosted or as Weaviate Cloud, with some enterprise features behind a licence key.

    Write a review
  6. C

    Chroma

    DataBase

    Chroma is an open-source embedding database for AI application developers, usable locally or as the hosted Chroma Cloud service.

    Write a review
  7. P

    pgvector

    DataBase

    pgvector is an open-source extension that adds vector storage and similarity search to PostgreSQL, so embeddings sit alongside relational data.

    Write a review
  8. L

    LanceDB

    DataBase

    LanceDB is an open-source, embedded vector database built on the Lance columnar format for AI and multimodal data retrieval.

    Write a review
  9. V

    Vespa

    DataBase

    Vespa is an open-source search and recommendation serving engine combining vector, text and structured retrieval, self-hosted or via Vespa Cloud.

    Write a review
  10. T

    Turbopuffer

    DataBase

    Turbopuffer is a proprietary, managed search engine offering vector, full-text and filtered search on an object-storage-native design.

    Write a review

For vendors

Want to see your product here?

The order is earned: ratings, verified reviews and listing quality decide it, and nobody can pay to move up. There are two ways to stand out, and each one says what it is.

Not on TrustList yet? Add your product or claim its listing, free.

Get vetted

Free

  • We check your registration, your website and your clients
  • Approved listings show the Verified badge here and on their profile
  • It earns a rank number and counts towards the TrustList Score; renewed yearly
Apply for vettingHow vetting works

Sponsor this ranking

Paid · always labelled

  • A Sponsored card above the ranked list, up to three per ranking
  • Marked Sponsored, and it never changes the order
  • Booked by the month; cancel any time
Sponsor this rankingSee prices

How is this ranking of vector databases worked out?

A position here is earned, never bought. It is worked out from what each listing documents and what has been verified about it: a complete, accurate profile, and the checks a vendor has passed. Client reviews count where they exist, but they are still rare across this catalogue, so today the order comes mainly from documentation and verification. A top position means best-documented, not best for your workload.

Is the top-ranked engine the best for everyone?

No. Workloads, scale, skills and budgets differ, and so does the right engine. Use the checklist above and test before you choose.

Can a vendor pay to move up?

No. Sponsored and Featured places are labelled and shown apart from the ranked list, and they never change an engine’s position in it. How rankings work

Articles

Latest insights

The newest guides and analysis from TrustList Insights.

All articles
Browse

More categories

Jump to another ranked category.

All categories
Keep exploring

More on TrustList

Everything here links back to the same verified catalogue. Pick your next stop.