Prowl Agent
What is wrong with my site that I can see from outside it — and did last week’s fix actually work?
How it goes
- 01
Type a domain
Nothing is spent and nothing is asked for. You see which report answers which question before any consent screen exists on the path.
- 02
Sign in with Google
So the run has somewhere to belong. It asks for your name and email address and nothing about your data — check it on Google's own screen.
- 03
Read the report, or hand it to an agent
The report has its own address. Under it is a block you can paste into Claude Code or any agent that can read a URL and edit files — it carries the report's address and no credential of ours.
- 04
Run it again next week
The second run says what changed: fixed, appeared, still there — and, separately, the findings on pages only one of the two runs opened, which are neither.
Three claims, and how to check each one
Every finding is one URL and one measurement.
Open the URL. If the page has a title, the finding was wrong and you can say so.
Every report names what it did not reach.
A crawl that read nine of forty pages says so, on the report, beside the nine — a smaller report, never a quieter one.
No model reads your pages.
Findings come from rules. Nothing about your site is sent to a model provider, and the privacy page says what is kept instead: a fingerprint of the text, not the text.
What it will not tell you
It does not promise traffic, positions or revenue, and it does not rank your site against anybody. A finding is a measurement; what it is worth is your judgement. The paid reports — what Google actually does with your pages, and what happens after the click — are not on sale yet, and their price will be on this site before anybody is asked for it.