Cloudflare Vectorize

What Cloudflare Vectorize is

Cloudflare Vectorize is a globally distributed vector database for building AI applications with Cloudflare Workers. It supports semantic search, recommendations, classification, anomaly detection, and supplying context to large language models. Vectors may come from Workers AI models or be provided by platforms such as OpenAI. Workers connect to an index through a Vectorize binding configured in Wrangler or the Cloudflare dashboard, and a REST API can insert and query vectors. Metadata filtering supports indexed string, number, and boolean properties, with up to 10 metadata indexes per Vectorize index. Each index can hold up to 20 million vectors, with a maximum of 1,536 dimensions per vector at 32-bit precision. Queries can return up to 50 results with values or metadata, or up to 100 without them. The Workers Free plan includes 30 million queried vector dimensions and 5 million stored vector dimensions per month. Cloudflare says it does not bill for CPU, memory, active index hours, index count, data transfer, or egress. Paid usage includes monthly allowances before additional dimension charges apply.

Who it suits

It suits developers building AI applications with Cloudflare Workers that need vector search, recommendations, classification, or model context. It may also fit teams whose embeddings come from Workers AI or services such as OpenAI.

Limits

Workers Free includes 30 million queried vector dimensions and 5 million stored vector dimensions per month, with 100 indexes per account and 1,000 namespaces per index. Workers Paid includes the first 50 million queried and 10 million stored vector dimensions, then charges $0.01 per million queried dimensions and $0.05 per 100 million stored dimensions; its plan allows 50,000 indexes per account and 50,000 namespaces per index.

Getting started

1. Open the Vectorize documentation website.
2. Choose the Workers Free plan or a paid plan.
3. Create an index and configure a Vectorize binding in Wrangler or the Cloudflare dashboard.
4. Connect from a Worker or use the REST API to insert and query vectors.
5. Use the #vectorize Discord channel for questions and feedback.

Verdict

Vectorize provides a Workers-connected database for vector indexing and querying, with defined capacity and query limits. The free plan includes monthly usage allowances; paid usage charges apply beyond the listed allowances.