Skip to content
Theyfetch

Theyfetch

Do they fetch you?

See whether AI assistants can read your pages, and which ones they fetch for live answers. Then see who they recommend instead.

Diagnose, fix, prove

Prompt tracking comes after they can reach the page. A 0% score is often a block, not a content problem.

  1. 01

    Diagnose

    We read robots.txt and fetch the homepage as ChatGPT-User and Claude-User. You see allow, block, or a challenge. Chrome can be fine while the assistant is not.

    Access check

  2. 02

    Fix

    Named actions, not a score. Allow Agent on Cloudflare. Stop Wordfence matching Claude-User. Server-render the article. We do not flip your WAF.

    What to change

  3. 03

    Prove

    After the change, ingest keeps verified ChatGPT-User and Claude-User hits. Training and search stay on their own rows. The empty state goes away.

    Pages they fetched

Start with access, pay for answers

Free crawler analytics and the AI access check. Prompt runs are on paid plans, every engine included.

Free

€0/ month

  • Crawler analytics
  • AI access check
  • 7-day citation history
  • 1 workspace
  • No prompt runs
Start free

Starter

€69/ month

  • 1,500 answers a month
  • ChatGPT, Claude, Perplexity, Google
  • 90-day history
  • 1 workspace
  • Weekly digest
Start starter

Growth

€199/ month

  • 6,000 answers a month
  • 3 workspaces
  • 1-year history
  • Weekly digest
Start growth

Scale

€499/ month

  • 15,000 answers a month
  • 10 workspaces
  • 1-year history
  • Weekly digest
Start scale

Answers are one prompt on one assistant, once. Every paid plan includes every launch engine. Cancel in the Stripe portal. Existing subscriptions stay on the plan you bought.

See pricing

Questions

What do you store?

Verified AI-bot hits, without the IP or the full user agent. Prompt answers, citations and who was named. We are the processor. You are the controller.

What is a live-answer fetch?

A request from ChatGPT-User, Claude-User, Perplexity-User or a similar token, made because a person asked. Training bots and search indexers are listed separately.

Why would Claude be blocked if Chrome works?

New Cloudflare zones can block Agent bots on pages with ads. Wordfence and some robots.txt lists also match Claude-User. That is the ticket we exist to catch.

Do you scrape ChatGPT or Claude?

No. ChatGPT and Claude runs use their APIs and are labelled api. Perplexity and Google AI Overviews come from their documented APIs. We do not drive the consumer apps.

What is on the free plan?

Crawler analytics, the AI access check, and 7 days of citation history for one workspace. Prompt runs start on paid plans.

Can I block training and still show up in answers?

Yes. Blocking GPTBot and ClaudeBot while allowing ChatGPT-User and Claude-User is a valid default. Blocking the user-fetch tokens and hoping to appear in ChatGPT is the contradiction.