Cloudflare AI Search Includes Free Allowances: Budget Before November Billing

October 3, 2026

News / Included usage and future costs

A small document-search pilot may fit included usage, but ingestion, storage, queries and answer generation need separate checks.

October 1, 2026: generally available; usage billing scheduled for November 1, 2026.

Included-usage containers sit beside separate meters for ingestion, storage, queries and answer generation.
AI-generated concept illustration. It does not show an official interface or a measured result.

Source information checked on October 3, 2026 at 07:56 (UTC+8). This brief uses official documentation and stated editorial calculations; the site has not tested the product or verified an individual account’s eligibility or bill.

Cloudflare made AI Search generally available on October 1 and says usage billing begins November 1. Its announcement lists monthly allowances including 10 GB of stored index data, 1,000 semantic queries and 1,000 full-text queries. Ingestion has a shared five-million-token allowance covering supported content. Official announcement and pricing table

For a small agency considering document search, those allowances provide a starting point for a limited pilot. Choose a small collection and estimate how often people will search it. Keep a record of failed searches as well as useful results.

The technical changelog says selected Workers AI embedding and reranking calls are included, while generation, query rewriting and external providers continue to use your account and gateway. A search allowance therefore does not establish that every generated answer is free. Official technical changelog

Estimate ingestion, storage and query usage before connecting a larger archive. Include maintenance time and review the account’s billing settings before November. Our automation cost guide helps turn the free allowance into a complete budget.

Limits and unverified details

The monthly included allowances are not a one-time promotion. Account conditions, index size, tokens and the actual bill are unverified. Excess usage and external models may add charges. No automatic stop at the allowance limit or universally free generated answers has been confirmed.