Setup guide

Choose a safe crawler-rendering setup and verify the response before changing your stack.

On this page

Start with the response

Before changing DNS, adding middleware, or installing a plugin, check the HTML that a crawler-style request receives. A page that already returns meaningful HTML does not need a rendering proxy. A page that returns only an application shell may need one.

Use a public crawler check to compare the initial HTML with the browser-rendered page. Record the hostname, status code, canonical URL, headings, and the first meaningful text before making changes.

Choose the smallest setup path

No-code DNS setup is appropriate when a hosted builder or static host cannot run server-side middleware. The public hostname is routed through a managed delivery layer while the existing host remains the origin.

Developer integration is appropriate when the application can run middleware, an Edge Function, a Worker, or server-side routing. It should route only eligible crawler requests and keep API credentials in server-side environment variables.

Keep the public and private boundaries clear

Only public, cacheable pages belong in a crawler-rendering path. Exclude account pages, dashboards, checkout, API endpoints, previews, password-protected content, and any route whose output depends on a session or cookie.

Never place an API key in browser JavaScript. Do not use a broad crawler user-agent match to expose content that ordinary visitors cannot access.

Verify after every change

  1. Test your homepage and one important public page, such as a product, service or article page.
  2. Record the starting URL, redirects and final HTTPS URL.
  3. Compare a normal response and a crawler-style response on that final hostname.
  4. Check the status, important text, links and canonical URL. Where rendering is configured, check response headers or delivery evidence too.
  5. Test root and www separately if both resolve.

A simulated crawler user agent is a diagnostic request, not proof of an authenticated bot visit or indexing.

A successful DNS record or an Active label alone does not prove that crawler requests reach the rendering path. The response is the proof.

If the response and setup status disagree, use the troubleshooting guide before changing DNS again.

Instructions for your connected site

Open account setup for domain-specific values, diagnostics and copy-ready configuration. These stay private to your workspace.