Cloudflare Browser Research

Cloudflare Browser Research

We expanded web research with a browser-backed fallback for pages that require more than a straightforward document request. This gives research a better chance of reaching useful public evidence while preserving a controlled execution path.

• Added browser-backed web retrieval for research scenarios that need page rendering or interaction.
• Kept browser execution behind the research workflow instead of exposing low-level tools to users.
• Improved source capture, metadata, citations, and evidence handoff from web retrieval into the final report.
• Added bounded timeouts, retries, provider fallback, and clearer partial-result states.
• Added live smoke tests and execution contracts for browser-backed research.
• Kept the user-facing experience focused on sources, findings, progress, and next actions rather than infrastructure details.

This makes research more resilient on the open web without turning the product into a collection of hidden technical controls.