OFICIAL Cloudflare Blog

Cloudflare AI Search: give your agents a search engine for your data

What happened
Based on Cloudflare Blog · Aug 06, 2026

Cloudflare launches AI Search to automate retrieval for agents, replacing manual integration of multiple AI tools. Early pricing preview emphasizes predictable costs, with free embedding and reranking for default models.

Cloudflare AI Search: give your agents a search engine for your data
Cloudflare Blog — Cloudflare
Key points
·
Today, the company is excited to announce a few developer experience improvements to Cloudflare AI Search to make it easy to manage a search solution out of the box.
·
Previously, you had to stitch together components of the Cloudflare primitives (Workers AI, AI Gateway, Vectorize, R2, Browser Run) but now, AI Search can do this automatically, and better.
·
Cloudflare its goal is to give your agents their own search engine, where they can easily find data to provide better answers for themselves and their humans.
·
Cloudflare is also sharing an early preview of pricing for customers of AI Search so you can learn how this scales.
Key numbers
·
For example, a 20,000-document source with 30,000 monthly queries is estimated at approximately $21 per month after initial indexing.

Cloudflare AI Search now automates the setup of a search engine for agents by integrating Workers AI, AI Gateway, Vectorize, R2, and Browser Run into a single service. Previously requiring manual stitching of components, the new tool handles crawling, ingestion, embedding, and retrieval automatically, reducing complexity for developers. It supports hybrid search, combining semantic and keyword matching to address both broad and specific queries. The service is designed to provide agents with direct access to relevant data, improving response accuracy without relying on external sources.

Pricing for AI Search is shared in an early preview, structured to ensure predictability and scalability. Embedding and reranking are free when using default models, eliminating concerns over token count variability. The preview outlines costs for ingestion, storage, and queries, with separate pools for tokens and query volume. For example, a 20,000-document source with 30,000 monthly queries is estimated at approximately $21 per month after initial indexing. Billing remains disabled during the beta period, with advance notice planned before activation.

AI Search is already deployed across Cloudflare’s own platforms, including Cloudflare.com, Developer Docs, and the Blog, with additional instances for tools like EmDash and the Cloudflare Dev Stack MCP. The Dev Stack MCP Server enables coding agents to retrieve current, cited documentation directly from the Cloudflare ecosystem, replacing slower methods like web searches. Users can deploy AI Search with a single command, configure public or private endpoints, and integrate it into existing applications or agents via MCP or custom domains.

The service leverages Browser Run’s /crawl for web scraping but operates under its own bot identity, Cloudflare-AI-Search, which respects robots.txt and site-specific bot controls. AI Search is available today at no cost during beta, with general availability pricing subject to change. Developers can enable it immediately, configure hybrid search, and deploy search endpoints for agents or end users. Documentation and setup guides are provided to facilitate integration.

Original source → Deals on Clipraptor.com →