Consumer View
The Weakest Line We Sell
One read a day of the answer a consumer is actually served, on ChatGPT, Google AI Overview and Microsoft Copilot, from $27 a month
What a daily read is
Once a day, on each surface you choose, we ask each of your prompts the way a signed-out consumer would and keep the answer we were served. Then we count: did it name you, did it cite your domain, which of your tracked competitors did it name. That is the whole measurement, and it is what most daily monitoring is.
What it can't tell you
- 01
One Read, Not Three
A single answer a day carries no range.
A report asks each question three times and prints how wide the answers wander. One read is one read: a line that moves may be the assistant changing its mind or the roll of the dice, and this page cannot tell you which.
- 02
Perplexity Held Out
A sign-up wall blocks the read today.
Four of four reads through Bright Data came back empty behind Perplexity's sign-up prompt, and the vendor names no date for that changing. So it is off the door: shown as held out, never sold, never charged for. The day a read gets through, it goes back into the line and this card goes.
- 03
The Model As Reported
Whatever the surface chose that morning, printed.
A consumer's ChatGPT decides for itself which model answers. We print the model the surface reported and never one we picked, so two mornings may not be the same measurement, and the page says so beside each answer. Where a surface names no model, the field says that instead.
- 04
Grounding As Found
Web search on or off, as the surface decided.
Where the surface tells us whether it searched, the flag is kept per answer. Where it does not, the field reads unknown rather than a guess.
- 05
The Answer Kept
Every count traces to the text we were served.
The stored answer sits on the page under the count it produced. A number that can't point at its answer isn't printed.
Why we sell it anyway
Of the 37 AI-visibility listings we read in the Shopify App Store on 12 September 2026, none stated a model version and two stated a sample count. This line states both, prints who reads each surface for us, keeps every answer, and costs $27: the price is on this page, not in a call. When you want the model named before you pay, every question asked three times and a range on every rate, that is the report, and the daily read is the cheapest way to find out you need one.
Build the report instead · the five marks it carries are the standard.