Public engineering tools
Can AI agents access your page?
Check crawler rules, delivered HTML, indexing controls and machine-readable links. Get the evidence and a list of fixes for one public URL.
A successful fetch is one step toward discovery. It does not promise that an assistant will find, cite or use the page.
Follow the access chain
- PermissionWhich robots rules apply to this exact path, separately for search, assistant and training controls?
- DeliveryWhat status, redirects, headers and HTML reach this checker? Do synthetic client declarations change the response?
- InterpretationIs text already in the HTML? Are canonical, indexing, structured-data and alternate-format signals available?
The checker does not run JavaScript or log into the site. A synthetic request with an AI bot’s name comes from this checker, not that provider’s network. Optional files such as llms.txt are reported as observed; their presence is no citation guarantee.
What does your request reveal?
Our analytics library keeps client declarations, browser-shaped headers, network evidence and verification separate. Inspect the request sent by this browser to see those distinctions. This uses one check from the same daily allowance.
Inspect this request as JSON. The JSON link examines a navigation request; the button examines its own fetch request. Their header shapes can differ. A declared name does not authenticate a provider or identify a person.
Use the free API
The same report is available to scripts and agents. No API key is required. The daily allowance is shared with the form above. Admitted checks count even when fetching fails.
curl https://gkoreli.com/api/ai-check \
-H 'Content-Type: application/json' \
-d '{"url":"https://example.com/article"}'
Request, response and quota contract · Source and tests
Reports describe the bounded requests made at the recorded time. Firewall rules, location, authentication and provider-specific behavior can change what another client receives. Use the result to choose the next test.