Authenticate the web page, not a screenshot of it

A paralegal pastes a URL into your evidence tool. Four GETs later, the capture exists.

GET /api/v1/browser/pdf?url=…&printBackground=true renders the page as it prints. GET /api/v1/browser/screenshot?url=…&fullPage=true renders it as it looked. GET /network?since=<start> on the same browser instance returns every request the page made: method, URL, status, tab: the part a screenshot can never show. All three land in the files service, and ?sha256 on each returns the 64-character digest for the capture log.

Nothing here required a plugin, a capture appliance, or an affidavit about a tool nobody can inspect. The capture is a set of HTTP responses from named endpoints, with hashes, on a server you rent. When opposing counsel asks how the page was preserved, you hand them the four URLs.