Pre-rendering Tester
See what each user agent is served: compare the HTML, status, title, content and links returned to crawlers and browsers. How it Works
Per User Agent
Status, title, canonical, robots and text for each crawler, against the browser of its kind.
Before and After JavaScript
Same HTML, a pre-render matching the rendered page, or different content: text similarity, not an exact hash.
Every Signal
Titles, canonicals, robots directives, links and structured data side by side.
A Real Example
Checked 17 November 2026https://www.pinterest.com/
- What the tool found
- 1 of 5 crawlers is refused: GPTBot (listed as OpenAI) gets a 403 titled "Pinterest - Forbidden". Googlebot Smartphone, Googlebot, Bingbot and Facebook get the same HTML as Chrome, a 1.20 MB shell with 0 words of text, sent with Vary: User-Agent; in Chrome on Android, JavaScript adds a little over 200 words.
- What to do
- Nothing to fix for search engines: no dynamic rendering, no cloaking, and Google renders the JavaScript. The GPTBot refusal is deliberate, since Pinterest's robots.txt disallows every crawler that isn't on its allowlist.
Reading the Result
Same HTML
The crawler gets exactly the HTML a browser gets, so there is no second version to keep in sync. If the browser rows show far more words after JavaScript, crawlers that don't render see only the shell.
Pre-rendered
The crawler gets its own HTML, and it matches the browser's page after JavaScript: dynamic rendering that works. Google treats dynamic rendering as a workaround, so plan a move to server-side or static rendering.
Different
The crawler's page matches neither the browser's HTML nor its rendered page: another status, title, canonical or meta robots, or under 90% of the text in common. Usually a stale or broken pre-render; serving crawlers other content on purpose is cloaking.
Refused
A 401, 403, 429 or 503 for that crawler only. It can be a deliberate block or bot protection that checks the IP address, which the real crawler would pass.
What it Can't Tell You
- Sites that choose the pre-rendered version by verified IP address rather than by user agent serve the real crawler a page this test never receives.
- Crawler responses are compared as raw HTML, without running their JavaScript. To see what a renderer makes of one user agent's page, use Fetch & Render.