Free AI Visibility Tools: Brand Report, AI SEO Audit, Query Fan-Out Generator, and Crawler Inspector

A pitch does not need a full prompt program. It needs one honest snapshot. We use the free tools on promptwatch.com for that. We do not pretend they replace Essential at $95/mo or Kick-off at $199/mo. The free set is a conversation starter. The paid project is where the work actually gets trended, diagnosed, and fixed. The distinction matters because a pitch is about earning the next meeting, not about proving a quarter of work we have not done yet.

Explore (free, 10 ChatGPT prompts) is the next step if the stakeholder wants to see a live project. The tools below do not require a seat. That matters in a pitch, because the procurement lead can open the same URL we are talking about and follow along. No login wall between us and the snapshot. When someone can click through and reproduce what we just showed, the conversation stays on the data instead of drifting into a debate about our methodology.

What we run before a retainer

AI Brand Visibility Report. Scores, sentiment, and competitive context across models. We use it to show that ChatGPT and Gemini already disagree. Two models, same brand, different answers. That single observation reframes the brief for most stakeholders. They stop asking for "an AI score" and start asking which model they care about. We do not treat the PDF as a baseline we will trend for a quarter. Trending needs a frozen prompt list and daily paid checks. A one-time PDF is a photograph, not a time series. The photograph is still useful. It tells you where the brand sits today, on the models the report covers, in the wording the report used. What it cannot tell you is whether that position moved, because there is no second check to compare it to.

AI SEO Audit. A pass on whether the site looks set up for ChatGPT, Claude, Perplexity, and Gemini. We pair it with a robots.txt read and, when we have access, a log check. The audit is a conversation starter. Agent Analytics on Professional, Business, or an agency plan is the proof. The audit tells you the site might be crawlable. The logs tell you the bot actually fetched the URL. Those are different statements, and we keep them separate in the room. A page can pass an audit and still never get fetched, because a CDN rule or a robots block sits between the audit and the crawl. Naming that gap is part of why the audit reads as honest rather than as a sales pitch.

ChatGPT Query Fan-Out Generator. Shows how ChatGPT expands a typed question into background searches. Useful in a workshop, because most stakeholders have never seen their buyer query fan out into the subquestions the model actually retrieves for. The paid product stores fan-outs per response, per model, on the prompts we freeze. One generator run is not a program. It is a demo of how the model thinks. We use it to pick better frozen prompts later, not to report on them. The fan-out view changes how people write prompts. They stop typing category labels and start typing the actual question a buyer asks in the middle of a decision.

AI Crawler URL Inspector. Extracted markdown and metadata as a model would see the URL. We use it when a page is a JavaScript shell and someone swears the claim is "on the page." The inspector shows what the model actually receives, which is often not what the browser renders. Token inspector sits next to it if the page is huge and we want to see how the text is chopped. Neither tool is crawl-to-citation. They tell you the page is readable. They do not tell you a model cited it. Readable is necessary. Cited is the outcome that pays. We keep those two words apart.

Also free, and we mention them when they fit: robots.txt generator for AI, the bots directory (400+ agents), and an llms.txt generator. We will not claim Google uses llms.txt for Overviews. If we generate the file, we call it optional documentation. A nicer about page for bots, not a ranking lever. The robots.txt generator is the one we reach for most, because a clean AI crawler allowlist is the kind of thing a client can ship in an afternoon and feel good about.

G2 rates the paid platform 4.7/5. That score is not a score for the free tools. We do not borrow the paid rating to dress up a free snapshot. The free tools stand on what they show, and the paid rating belongs to the paid product.

What the free set cannot do

It cannot hold 50 prompts across ChatGPT, Gemini, Claude, Perplexity, Grok, Llama, DeepSeek, Mistral, Copilot, AI Overviews, and AI Mode on a daily refresh. That is the operational gap. A retainer needs a frozen list, refreshed on a schedule, with the same wording each check. The free set does not store that. Without a frozen list, week one and week two are not comparable, and a trend is just two unrelated numbers.

A Professional, Business, or agency project can join crawler errors to citations. The free set and Essential cannot. This is the diagnostic gap. When a citation drops, the paid project lets us ask whether the bot stopped fetching the page before we ask whether the copy changed. The free set has the citation side or the crawl side, never the join between them. That join is where most of the real work happens. A citation that disappeared the same week the CDN started returning 403 to ClaudeBot is a different ticket from a citation that disappeared after a rewrite.

The free set also cannot open a review inbox into Webflow or Framer. WordPress publish is still coming soon on the paid CMS side anyway. So even the paid Content Agents loop is not fully closed on WordPress today. We do not promise it in a pitch. Promising a feature that is not shipping yet is the fastest way to lose a retainer in month two.

Otterly at $29 is a paid mention checker with Gemini as an add-on and lag. Peec at $95 is paid. Profound Starter at $99 annual is ChatGPT only. A free Promptwatch report is still the cheaper way to start the conversation. It is not those subscriptions, and it is not Kick-off. We name them so the stakeholder knows the free snapshot sits in a real market, not a vacuum. Naming the neighbors also keeps us from overselling. The free report does what it does, and the paid neighbors do what they do.

Professional is $245/mo and is the first brand plan with crawler logs. The agency plans include crawler logs too. Essential has no listed crawler-log allowance. Business is $579/mo. We do not quote a $29 Promptwatch plan. We do not promise instant alerts from a free generator. The free tools run on demand. They do not push to your phone. On demand is fine for a pitch. A phone push is a different product, sold on a different plan.

What we will not do in a pitch

Invent a client result on top of the brand report. Screenshot one ChatGPT tab and call it an audit. Ship an llms.txt as the entire technical recommendation. Autopublish anything.

These are the four ways pitches go wrong with free tools. Each one looks like progress and is not. A fabricated result poisons the trust the snapshot was supposed to build. A single screenshot is one model on one day, which is the opposite of an audit. An llms.txt file is a document, not a strategy. Autopublishing from a free generator skips the human review that makes content safe to ship. The pattern across all four is the same: a free tool is stretched past what it actually does, and the stretch is what breaks the engagement.

If the audit says crawlers are blocked, that is the first paid-week ticket. See how to get cited by ChatGPT. A blocked crawler is a concrete, fixable thing, which is exactly the kind of finding a pitch should end on.

FAQ

Are the free tools enough for a monthly GEO retainer?

No. They are enough to decide whether a retainer is worth opening. The retainer needs a project, a frozen list, and someone who will reject bad drafts. The free set gives you the photograph. The retainer pays for the time series and the editorial loop.

Can the client run the tools without us?

Yes. The URLs are public. We still prefer to run them together so the prompt in the fan-out generator is a real buyer question, not "best software." A fan-out run on a generic prompt teaches nobody anything about the brand.

How we run it

1. Pick five buyer prompts. Run the fan-out generator on two of them. We want the stakeholder to see the model expand a real question, not a category label.
2. Run the brand report and the AI SEO audit on the production host. Not the staging domain. The model fetches production, so we inspect production.
3. Inspect one money URL in the crawler inspector. Compare to the rendered page. If the inspector markdown is missing the claim, the audit finding writes itself.
4. If the snapshot is enough to start, open Explore or Essential for prompt tracking. Use Professional when a brand project needs crawler logs, or Kick-off when we are holding several logos. The plan choice follows the scope, not the other way around.
5. hello@1001seomedia.com if you want us to run the snapshot and then the program. Send the five prompts. Do not send a promised score.