Redirect action (network request layer)

The Redirect action rewrites matching requests at the browser’s network request layer — the original destination never receives them. Two modes are available.

Replace origin

Swaps scheme, host and port of the target while keeping the original path and query — https://api.prod.com/users?page=2 becomes http://localhost:3000/users?page=2. The target may carry a path prefix: the original path is appended after it.

The query-params table can add, replace or remove parameters on the redirected URL — handy to append env=test everywhere, or to strip a cache-buster like _t.

Regex substitution

Replaces the first fragment of the URL that matches the pattern. To replace the whole URL, anchor it yourself with ^ and $, and reference capture groups with \1, \2… The replacement must be a whole http(s) address.

Match:        ^https://api\.example\.com/(.*)$
Replace with: https://api.test.example.com/\1

Caveats

• The browser treats the redirected request as a real cross-origin request — CORS may apply. Add a Headers rule to supply Access-Control-Allow-Origin.
• Cookies follow the target domain, so endpoints that need a login session may fail.
• An https page can only redirect to https or localhost — intranet http targets are blocked by the browser.
• Responses produced by the page's own Service Worker or cache bypass the browser's network request layer.
• Avoid rules that keep rewriting their own output (rules pointing at each other, or substitutions that grow the URL) — that forms an endless request loop. A target that resolves to the current URL is safe: the browser skips a redirect that changes nothing.

The panel detects such loops before they start: a chain of redirect rules that keeps rewriting the same URL is flagged on the rule, with the chain shown in the warning.