Does Puppeteer Follow Redirects Automatically When Request Interception Is Enabled?

Answer

When page.setRequestInterception(true) is active, Puppeteer does not transparently follow HTTP redirects. Each 3xx response generates a new request event for the Location URL, and the original request receives the redirect response. You must explicitly allow or abort each hop.

Typical Flow

1. Original request is paused.
2. Browser receives a 301/302/303/307/308 response.
3. That response is delivered to the request event listener.
4. You call request.continue() to let the request proceed.
5. A second request event is fired for the redirect target URL.
6. The process repeats until a non‑redirect status is returned.

What You Must Do

• In the request handler, call request.continue() for every request, including those that are part of a redirect chain.
• If you want to stop the navigation, use request.abort() on the target redirect request.
• To inspect the Location header, read request.headers() or the response’s response.headers() after request.continue() resolves.

No Built‑In “Keep Redirects Transparent” Option

The interception API deliberately exposes every network request, including those generated by redirects. There is no flag to hide redirects while keeping interception enabled.

Practical Workarounds

• Disable interception for the navigation itself:

  await page.setRequestInterception(false);
  await page.goto('https://example.com');
  await page.setRequestInterception(true);

  This lets Chrome follow redirects automatically, then you re‑enable interception for subsequent sub‑resource requests.
• Use CDP directly to filter out redirect requests if you only care about certain URLs:

  const client = await page.target().createCDPSession();
  await client.send('Network.setRequestInterception', {
    patterns: [{ urlPattern: '*', requestStage: 'Response' }]
  });

  But this still gives you a requestIntercepted event for each redirect.

Example Listener

page.on('request', request => {
  const url = request.url();
  console.log('Request:', url);
  // Let every request proceed – redirects will trigger new events
  request.continue();
});

Diagnostic Question

Do you need to preserve the original navigation state (e.g., keep the page at the final URL) while mocking only API calls, or do you want to mock every request including redirects? This determines whether the “disable‑then‑enable” pattern is appropriate.

Verification Steps

• Turn on interception.
• Navigate to a URL that returns a 302 to a known endpoint.
• Log each request.url() in the event handler.
• Confirm you see two events: the original URL and the redirect target.