Connect your website

Open Install snippet in your workspace. Add your site and copy its ingest token when it is shown.

Keep the token in your server configuration. It is shown once; if you lose it, rotate it in the workspace.

Any server that can POST JSON can use the HTTP endpoint. Forward the original page request’s method, URL, IP, and user agent.

curl -X POST https://theyfetch.com/api/ingest \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "timestamp": "2026-09-02T10:00:00Z",
    "method": "GET",
    "url": "https://example.com/blog/post",
    "ip": "203.0.113.5",
    "geo": { "country": "GB" },
    "referer": "https://chatgpt.com/",
    "userAgent": "Mozilla/5.0 ... ChatGPT-User/1.0 ...",
    "status": 200
  }'

These values illustrate the request shape. Use the real request’s timestamp and metadata in your integration. The example IP is not a provider IP and will not produce a verified read.

1. Deploy the server integration and open Install snippet in your workspace.
2. Watch the last-fetch status. Human page visits do not populate the AI analytics.
3. Run the AI access check to find blocked fetchers, challenge pages, or HTML an assistant cannot read.

Check that your integration runs on public GET requests. If a firewall blocks the assistant before it reaches your server, the snippet cannot record that attempt.

Use the access check to inspect your host and robots.txt settings. An open page can still have no reads if an assistant has not chosen to fetch it.