# Which vector database works best for AI agents?

Of the 7 vector databases we drove up to 2 Nov 2026, LanceDB, Qdrant, Zilliz and Weaviate work best for AI agents. With them, our agent got in, did nearly all the core work and could leave without a person. Sign-up stopped our agent at 2 of 7.

Source: https://agentdrives.com/research/best-vector-databases-for-ai-agents?via=md · Published 2026-11-03

- Ranked: [7 vector databases of the 246 products we drove](/data/facts#cat.vector-databases.n)
- Driven: 26 Oct to 2 Nov 2026, signed agent
- Order: Level, then core work, then minutes to get in
- Main step: Get in. Also covers: Find, Use, Pay, Leave
- Data: [CSV](https://agentdrives.com/data/best-vector-databases-for-ai-agents.csv?via=md) · [JSON](https://agentdrives.com/data/best-vector-databases-for-ai-agents.json?via=md)

## Which one for which job

7 vector databases, driven 26 Oct to 2 Nov 2026.

- **[LanceDB](/tools/lancedb)** (L4): Best for vectors stored in the agent's own process. The open-source library installed from npm and PyPI and ran every core task with no account and no key: tables, upserts, deletes, vector, full-text and filtered search, indexes and branches. 8 of 8 core tasks. In after 2 min. Driven 30 Oct.
- **[Qdrant](/tools/qdrant)** (L4): Best for a hosted vector database an agent runs by API. Sign-up took an email code, and our agent made one management key in the web console. 7 of 7 core tasks. In after 4 min. Driven 2 Nov.
- **[Zilliz](/tools/zilliz)** (L4): Best for an agent that wants a CLI and a REST API for everything. After a sign-up in the browser, clusters, collections, inserts, full-text and hybrid search, imports and access control all worked over REST, the CLI and the SDK. 7 of 7 core tasks. In after 6 min. Driven 28 Oct.
- **[Weaviate](/tools/weaviate)** (L4): Best for answers over the data, not only search. After a sign-up with an email code, our agent made a free cluster and a key in the console, then did every core task without a browser: objects, vector, keyword, hybrid and filtered search, Query Agent answers with sources and Engram memories. 10 of 10 core tasks. In after 8 min. Driven 28 Oct.

## The ranking

| # | Vector database | Level | Find | Get in | Use | Pay | Leave | Core work | In after | Driven |
| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
| 1 | [LanceDB](/tools/lancedb) | [L4 Runs on its own](/method/levels#l4) | On its own | On its own | On its own | Stopped | On its own | 8 of 8 | 2 min | 30 Oct 2026 |
| 2 | [Chroma](/tools/trychroma) | [L1 Needs a person](/method/levels#l1) | On its own | Stopped | Not reached | Not reached | Not reached | Unknown | Unknown | 31 Oct 2026 |
| 3 | [Pinecone](/tools/pinecone) | [L1 Needs a person](/method/levels#l1) | On its own | Partly | Not reached | Not reached | Not reached | Unknown | Unknown | 26 Oct 2026 |
| 4 | [turbopuffer](/tools/turbopuffer) | [L1 Needs a person](/method/levels#l1) | On its own | Stopped | Not reached | Partly | Partly | Unknown | Unknown | 31 Oct 2026 |
| 5 | [Qdrant](/tools/qdrant) | [L4 Runs on its own](/method/levels#l4) | On its own | On its own | On its own | Partly | On its own | 7 of 7 | 4 min | 2 Nov 2026 |
| 6 | [Zilliz](/tools/zilliz) | [L4 Runs on its own](/method/levels#l4) | On its own | On its own | On its own | Partly | On its own | 7 of 7 | 6 min | 28 Oct 2026 |
| 7 | [Weaviate](/tools/weaviate) | [L4 Runs on its own](/method/levels#l4) | On its own | On its own | On its own | Partly | On its own | 10 of 10 | 8 min | 28 Oct 2026 |

Based on 7 of the 246 products we drove, last driven 2 Nov 2026.


## What our drives found

7 vector databases, driven 26 Oct to 2 Nov 2026.

- Minutes until our agent was in: LanceDB 2 min, Qdrant 4 min, Zilliz 6 min, Weaviate 8 min. Didn't get in on its own: Chroma, Pinecone and turbopuffer.
- Got in on its own: [4 of 7](/data/facts#cat.vector-databases.access_own). A key with no person: Qdrant, Weaviate and Zilliz.
- Has an MCP server: [7 of 7](/data/facts#cat.vector-databases.mcp). LanceDB, Pinecone, Qdrant, Chroma, turbopuffer, Weaviate and Zilliz.
- Paid on its own: [0 of 7](/data/facts#cat.vector-databases.pay_own). An agent payment on offer, tested or documented: Chroma.
- Left on its own: [4 of 7](/data/facts#cat.vector-databases.leave_own). LanceDB, Qdrant, Weaviate and Zilliz.

## Step by step

7 vector databases, driven 26 Oct to 2 Nov 2026.

- **Find:** [7 of 7](/data/facts#cat.vector-databases.find_own) had docs it could use. LanceDB, Pinecone, Qdrant, Chroma, turbopuffer, Weaviate and Zilliz. With a CLI: Pinecone, Qdrant, Chroma and Zilliz.
- **Get in:** [4 of 7](/data/facts#cat.vector-databases.access_own) got in on its own. LanceDB, Qdrant, Weaviate and Zilliz. First key made in a browser: Qdrant, Weaviate and Zilliz.
- **Use:** [4 of 7](/data/facts#cat.vector-databases.use_own) did the work on its own. LanceDB, Qdrant, Weaviate and Zilliz. Runs on its own (L4 or higher): LanceDB, Qdrant, Weaviate and Zilliz.
- **Pay:** [0 of 7](/data/facts#cat.vector-databases.pay_own) paid on its own. None of the 7. First stop: payment: LanceDB, Qdrant, Weaviate and Zilliz.
- **Leave:** [4 of 7](/data/facts#cat.vector-databases.leave_own) left on its own. LanceDB, Qdrant, Weaviate and Zilliz. Not reached: Chroma and Pinecone.

## Database by database

### [LanceDB](/tools/lancedb)

L4 Runs on its own. Find: On its own · Get in: On its own · Use: On its own · Pay: Stopped · Leave: On its own.

**Best for vectors stored in the agent's own process.** The open-source library installed from npm and PyPI and ran every core task with no account and no key: tables, upserts, deletes, vector, full-text and filtered search, indexes and branches. The hosted side is closed to agents: LanceDB Cloud takes no new sign-ups, and its REST API comes with Enterprise, sold through a sales form.

### [Chroma](/tools/trychroma)

L1 Needs a person. Find: On its own · Get in: Stopped · Use: Not reached · Pay: Not reached · Leave: Not reached.

Chroma is a database that stores data for AI search. The agent read all the public docs and specs, but could not make an account. The sign-up answered 'sign-ups are temporarily disabled' every time, and Chroma says its cloud service is winding down.

### [Pinecone](/tools/pinecone)

L1 Needs a person. Find: On its own · Get in: Partly · Use: Not reached · Pay: Not reached · Leave: Not reached.

Pinecone is a vector database that stores data so software can search it by meaning. Our agent could read the documentation and list the hosted models on its own, but it could not create an account. The email sign-up showed a Cloudflare 'Verify you are human' checkbox.

### [turbopuffer](/tools/turbopuffer)

L1 Needs a person. Find: On its own · Get in: Stopped · Use: Not reached · Pay: Partly · Leave: Partly.

turbopuffer is a database for searching text and vectors. An agent could create an account with an emailed code, but could not get an API key because the key sits behind a paid card subscription with no free plan. The agent got as far as the card form and did no database work.

### [Qdrant](/tools/qdrant)

L4 Runs on its own. Find: On its own · Get in: On its own · Use: On its own · Pay: Partly · Leave: On its own.

**Best for a hosted vector database an agent runs by API.** Sign-up took an email code, and our agent made one management key in the web console. Everything after that ran without a browser: it created a free cluster over the Cloud API and did every core task, server-side embeddings included. On the hosted cluster, filtered search and delete-by-filter first need a payload index.

### [Zilliz](/tools/zilliz)

L4 Runs on its own. Find: On its own · Get in: On its own · Use: On its own · Pay: Partly · Leave: On its own.

**Best for an agent that wants a CLI and a REST API for everything.** After a sign-up in the browser, clusters, collections, inserts, full-text and hybrid search, imports and access control all worked over REST, the CLI and the SDK. The documented API key endpoints answered with errors, so the first key comes from the console or a device login confirmed in a browser.

### [Weaviate](/tools/weaviate)

L4 Runs on its own. Find: On its own · Get in: On its own · Use: On its own · Pay: Partly · Leave: On its own.

**Best for answers over the data, not only search.** After a sign-up with an email code, our agent made a free cluster and a key in the console, then did every core task without a browser: objects, vector, keyword, hybrid and filtered search, Query Agent answers with sources and Engram memories. New free clusters refuse the deprecated GraphQL API, which its docs still show in places.

## How we drove them

Our agent drove each vector database between 26 Oct and 2 Nov 2026: find the docs, get in, do the core work, pay, leave. It signs every request, never solves a CAPTCHA and signs up with its own address. [Our method](/method).

## How we order them

Level first, then the share of core work done, then minutes to get in. Levels run from L0 to L5. [What the levels mean](/method#levels).

## Limits

- One drive per vector database, its newest. A second drive confirms a level.
- Not in this sample: a memory layer built for agents.

## Sources

- [Qdrant MCP server](https://github.com/qdrant/mcp-server-qdrant)
- [Zilliz MCP server](https://github.com/zilliztech/zilliz-mcp-server)
- [Weaviate MCP server](https://docs.weaviate.io/weaviate/configuration/mcp-server)
- [Every number and the products behind it](/data/facts)

## Questions

### Can an AI agent sign up for Pinecone on its own?

No. A Cloudflare Turnstile check kept our agent out at sign-up, so it never had an API key. Without one it could only search Pinecone's docs and list its hosted models.

### Which vector databases have an MCP server an agent can use?

All 7 publish one. Our agent did real work through the servers of Qdrant, Zilliz and Weaviate. LanceDB's only searches its docs, and the servers of Chroma, Pinecone and turbopuffer need an API key it couldn't get.

### How much does a vector database cost at checkout?

At the checkouts our agent reached: turbopuffer $16 a month, Qdrant $26.10 a month, Weaviate $45 a month and Zilliz $99 a month, each a card form. Qdrant, Weaviate and Zilliz have a free plan before that; turbopuffer has none. LanceDB sells its hosted plan through sales only.

### Can an AI agent pay for a vector database on its own?

Not yet: 0 of 7. Each takes a card on a checkout form, and none offers an agent payment such as x402.
