Find out what an AI agent sees when it reaches your site

Agents do not browse. They look for the files a site publishes about itself, and most sites publish none. Measure yours in about eight seconds, no account.

Control what they read. Get ahead of what they say.

  • ChatGPT
  • Claude
  • Grok
  • Gemini
  • DeepSeek
  • Perplexity
  • Copilot
  • Meta AI
  • Mistral

The proof

Proof, not promises.

Anyone can score your markup. We send a real agent to run a customer's errand on your site and hand you the replay, so you see it work before your customers do, and publish the proof when it makes you proud.

How an agent task works

agent task · ecommerce.find_product_price
$ run task on example.com

reading /products/atlas-lamp
checking the price against the page
checking stock wording

success in 7 steps
"The Atlas lamp costs $89 and ships today."

replay recorded · published at the owner's discretion

Share of answer

Then ask the assistants what they say about you.

Your buyer asks an assistant before they open a search box, and the assistant answers with a handful of names. We put your category's real questions to the engines with web search on, keep every answer word for word, and count who was cited: you, a rival, or somebody you have never heard of.

Asked 47 times across 3 engines on September 6, 2026, not one answer named axrush.com. 4 other domains were named instead. We publish that the way we publish our own score.

How share of answer works

share of answer · one pass, one engine
$ ask anthropic · openai · gemini

"Which tools check whether a site is ready for AI agents?"

cited   1 cloudflare.com
        2 isitagentready.com
        3 agentchecker.ai

you     not cited, not mentioned

answer stored word for word · sources kept in order

The loop

Measure, watch, fix, prove. One loop, kept short.

Measure

Every page template of your site, on named checks with public weights. A score that means the same thing everywhere it lands.

Watch

History per check, scheduled rescans, drift disclosures, and alerts on rules you write. A score is a photograph; the series tells the truth.

Fix

A guide per check, files written from your own pages, one-click autofix, and prompts your AI pair can run. Grounded, never invented.

Prove

A real agent runs a customer's errand on your site. The replay, verdict, steps and hallucinations included, is yours to publish.

Developers

The same measurements, from your pipeline

An anonymous API and an MCP server for the public surface; an organization key for your own sites, with a CI gate in two curls. The gate lives in your pipeline, not in a browser tab.

Read the documentation

axrush · api & mcp
$ curl -X POST https://axrush.com/api/v1/scan \
  -H "content-type: application/json" \
  -d '{"domain":"example.com"}'
{ "host": "example.com", "score": 87, "grade": "good", … }

$ claude mcp add --transport http axrush https://axrush.com/mcp
added mcp server: axrush

See what the machines see. It takes eight seconds.