Cloudflare Browser Run snapshot adds four formats
Cloudflare's Browser Run updated its /snapshot endpoint to extract PDFs, links, scraped DOM elements, and AI-structured JSON in a single browser navigation. Developers no longer need to reload pages multiple times to collect visual captures alongside extracted text and structured schemas.
Running duplicate headless browser sessions just to extract visual artifacts and structured JSON is an expensive anti-pattern for web agents. Packaging full-page rendering and Workers AI extraction into a single edge call significantly reduces latency and compute costs for automated scrapers.
- –Cuts edge compute overhead by sharing a single Chromium session across visual, textual, and structured formats
- –Embeds Workers AI directly into the capture workflow for schema-validated extraction without secondary inference calls
- –Employs an all-or-nothing execution model within a 60–120 second timeout where any format failure aborts the entire request
- –Requires specifying at least two formats, directing single-target jobs to dedicated lightweight endpoints
DISCOVERED
1h ago
2026-11-20
PUBLISHED
1h ago
2026-11-20
RELEVANCE
AUTHOR
CloudflareDev