Technical Layer · Generator·Updated Aug 2026

Bot-Fetch Test Command Generator

The only valid rendering test is a raw HTTP fetch with no browser. Enter a URL and a content string and get a ready-to-run curl test and a five-bot sweep.

Your page and a content string
URL to test
A string only your real content has

Generates a single-URL curl test and a multi-bot sweep. Run them from any terminal. If the string is missing from the raw HTML, it is missing from the model's view of your page.

Your test commands
Enter a URL and a content string.

Testing the way a bot fetches, not the way you browse

The only valid rendering test is a raw HTTP fetch with no browser, run against the URLs that carry commercial weight. This generator writes two: a one-line curl that greps for a string only your real content has, and a sweep that tests the whole AI-crawler fleet at once. The word-count column in the sweep is the tell, 40 words to every bot and 1,400 in a browser and you have found your ceiling.

How to use it

  1. Enter the URL you want to test, usually your pricing page first.
  2. Enter a string that only exists in the real content, a price, a plan name or a specific feature.
  3. Copy the single-URL curl to check one page, or the sweep to test the fleet.
  4. Run it from any terminal. No output from the grep means the string is not in the raw HTML, which means it is not in the model's view of your page.

Why test with curl instead of my browser?

Because your browser runs JavaScript and the AI crawlers do not. Opening the page in Chrome shows you the rendered version; a raw curl with a bot user agent shows you exactly what a non-rendering crawler receives, which is the view that decides whether you get cited.

What user agents does the sweep test?

GPTBot, OAI-SearchBot, ChatGPT-User, ClaudeBot and PerplexityBot, the fleet that covers OpenAI, Anthropic and Perplexity. Different bots occasionally get different treatment because of an old CDN bot rule, which is exactly why testing the fleet beats testing one agent.

What does the word-count column tell me?

It is the fastest signal of a rendering gap. If a page returns 40 words to every bot but 1,400 in a browser, its content depends on a runtime the bots do not run. A healthy server-rendered page returns a similar word count to bots and browsers alike.

Is anything sent to a server?

No. The commands are generated in your browser and run on your machine. This tool uploads nothing.