Promptwatch vs VisiScan 2026
The platform we run client GEO on versus a four-engine local scanner. What we would buy, and why $29 Monitor is not our Promptwatch seat.
We get this comparison when a local owner ran VisiScan for free and asks why the retainer still includes Promptwatch. Short answer: a 40-answer snapshot is not a program. We use Promptwatch on client work. We might attach a $49 VisiScan page to a trades audit. We do not treat weekly 12 questions as the KPI source.
The question is fair, so we walk through it on the call. A free scan that returns 40 answers across four engines feels like evidence. It is evidence of one moment on one URL. It is not a log of what ChatGPT said on Tuesday, what Claude said on Thursday, which bot fetched the page, and whether the visit converted. Those four things are the program. The snapshot is the snapshot. Conflating them is how a client ends up with a deck full of grades and no record of what actually moved.
How we compared them this week
VisiScan site, pricing.md (24 August 2026), about, llms.txt, 29 August 2026. Promptwatch from the sheet (23 August 2026). We did not treat Harborlight Plumbing as a customer. We did not recount 328 scans.
We are explicit about what we read and what we skipped, because the comparison lives or dies on the receipts. The VisiScan pricing page and the pricing.md file disagree in places, and we note where. The Promptwatch numbers come from the fact sheet we keep sourced from promptwatch.com/llms.txt. Harborlight Plumbing is the demo plumber on the VisiScan site, not a client we can call. The 328 scans figure is a vendor stat about their own index, not a benchmark we would put in a client deck.
Mike Holp. Four engines. 1 free scan per URL.
Four engines means ChatGPT, Claude, Perplexity, and Gemini. One free scan per URL is the cap on the free tier, and it matters more than it looks. A client who reruns the same domain tomorrow to check progress hits that wall. The free pass is single-use per URL, not a standing meter.
What we would put in a client deck
| Question the client asks | What we answer |
|---|---|
| Did ChatGPT name us this hour? | VisiScan free, four engines, 5 questions. Promptwatch Explore if we only need ChatGPT prompts on a standing meter. |
| Can we host a weekly grade for $29? | VisiScan Monitor. We still need Promptwatch for logs and conversions. |
| Who publishes to Webflow? | Promptwatch Content Agents. VisiScan pastes schema. |
| Who sees GPTBot in CDN logs? | Promptwatch. VisiScan checks crawler access as a signal. |
| Is $29 Promptwatch? | No. Promptwatch Essential is $95. There is no $29 Promptwatch plan. |
We walk the table top to bottom on the call. The first row is the cheap question, did the model name us, and the cheap answer is either tool at the free tier. The second row is where the $29 Monitor pitch lives, and our answer keeps Promptwatch on the invoice because a weekly grade does not show logs or conversions. The third row separates a publish loop from a paste. The fourth row separates a CDN log from a readiness checkbox. The fifth row is the one we stop on the call, because someone has heard there is a $29 Promptwatch plan and there is not.
Email-versus-$49 copy goes in a footnote. We show pricing.md.
The email-versus-$49 fight is a copy conflict on one homepage block, and we do not resolve it for the client. We show the pricing.md file and let them read it. The footnote keeps the deck clean.
What we would actually buy
Internally: Promptwatch stays. Free VisiScan on a local inbound lead if we want a named-rate screenshot in the first email.
For a clinic that will never log into a GEO cockpit: $49 once, we still recommend Promptwatch Explore so ChatGPT is sampled on a real tracker if they grow.
For us-shaped retainers: Promptwatch Essential or Professional. VisiScan is optional appendix.
We would not sign a SOW that uses VisiScan's agent ladder as proof we shipped GEO content. Paste is not publish. The ladder is a readiness grade (L1 structured data, next rung machine-readable services, in their example). We already write pages. We do not need a $49 file to know a clinic should add LocalBusiness schema. If the owner will not log into Promptwatch, we still do.
The ladder is a site readiness audit, not a content calendar. L1 is structured data present. The next rung is machine-readable services. That is useful as a technical checklist. It is not a substitute for the publish, fetch, cite loop. We will not let a readiness grade stand in for shipped GEO content in a statement of work.
pricing.md caps the free scan at one per URL. A client who reruns the same domain tomorrow may hit that wall. Explore Promptwatch does not. We mention that before they burn the free pass on a staging host.
This is the small print that changes behavior. The free pass is one per URL. A staging host is a URL. If the client spends the pass on staging, they have none left for production. Explore has no per-URL cap, so it survives a rerun. We say this before they click, not after.
The recommendation we give
If they are choosing one tool because both say AI visibility, choose Promptwatch for anything that is a brand program. Choose VisiScan free for a local flashlight. If they already paid $49, keep the PDF and add Promptwatch anyway.
Full reports are one purchase per scan. A second location is another $49. Monitor is per the billing portal, weekly, still four engines. We would rather put that money on Essential once they have more than a dozen real buyer prompts. Twelve generated questions from a homepage are not a prompt strategy.
Twelve generated questions are a guess at what buyers might ask. A dozen real buyer prompts are what sales actually heard. The first is a content exercise. The second is a measurement list. Essential is built for the second. Monitor is built for the first. That is why the money goes to Essential once the real prompts exist.
That is the comparison we will sign. VisiScan is a local scanner. Promptwatch remains the measurement answer on our accounts.