Retrieval Window Calculator
See how much of your page ChatGPT actually reads versus what JavaScript hides from it.
Why we built it
ChatGPT's fetcher reads the raw HTTP response and skips JavaScript execution. Anything a page renders client-side sits outside that retrieval window, invisible to the model even though it looks fine in a browser. Few teams check how much of their content falls on the wrong side of that line, because checking by hand means reading view-source. So we are building the diff.
How it runs
- 01
Fetch the raw response
A plain HTTP request pulls the page the way an AI fetcher does, with no JavaScript execution, and the workflow extracts every text byte present in that response. This is the retrieval window.
- 02
Render the full page
Headless Chrome loads the same URL, executes the JavaScript, and extracts the text a human visitor ends up seeing. The gap between the two passes is the content the model can never read.
- 03
Diff and score
A line-by-line visibility map marks each passage as readable or hidden, and the retrieval score states what share of the rendered content survives in the raw pass. Low score, clear fix list.
Want this running on your stack?
We build, calibrate, and hand over each system to your team. Workflows export as JSON. Dashboards live in your accounts. No lock-in.
Related systems
AI Bot Verifier
Bulk paste log IPs. Get back which are real AI crawlers and which are impersonators.
AI Citation Tracker
Track how often ChatGPT, Perplexity, and Claude cite your brand for the queries that matter.
HCU Content Checker
Paste any URL. Score it against Google's Helpful Content signals in 30 seconds.