What a ChatGPT rank tracker does
A Google rank tracker reads a list that already exists: ten results, in an order Google chose, and your URL somewhere in it or not. An AI rank tracker has nothing to read until it asks. It sends the engine a question, waits for the answer to be written, and then works out which brands the answer named and in what order.
That is what this page does for one question at a time. You type what a buyer would ask, such as “best CRM for a 10-person agency”, plus your brand name, and optionally your domain and up to 5 competitors. The checker asks ChatGPT and Gemini, finds every brand and product each answer names, and orders them by where they first appear in the text. You see your position in that answer, the names ahead of you, the sentence that describes you, and the full answer to read.
If you have searched for an LLM rank tracker or an AI rank checker, this is the same job. The difference is in what we are willing to call the result, which is the next section.
Why there is no stable ChatGPT ranking
An answer from ChatGPT is generated, not retrieved, and it varies a lot. SparkToro and Gumshoe had 600 volunteers run 12 prompts through ChatGPT, Claude and Google’s AI, 2,961 runs in all. There was less than a 1 in 100 chance that ChatGPT or Google’s AI would give the same list of brands in any two responses, and less than 1 in 1,000 of the same list in the same order[1].
Presence is steadier than position. In the same study, City of Hope appeared in 69 of 71 ChatGPT answers but was the first name in only 25 of them[1]. Rand Fishkin’s conclusion was that a visibility percentage across many prompts, run several times, is a reasonable metric, and that “any tool that gives a ‘ranking position in AI’ is full of baloney”[1].
The variation is not total, though. SISTRIX tracked 82,619 prompts weekly for 17 weeks and found that “for 86% of all prompts, there is a stable core comprising a few domains; the rest rotates at a rate of 89% per week”[2]. SE Ranking found that only 9.2% of cited URLs matched across three runs of the same query in AI Mode on one day[3]. Some places in an answer are held. Most are passed around.
How this checker works
Two limits are worth knowing. First, these are the engines’ budget API models, the same ones our free AI visibility checker uses, because the larger models cost several times as much per answer. Second, they answer without searching the web. The ChatGPT app can search, and when it does it often rewrites your question into more specific queries before answering[4]. In the API, web search is a tool the developer has to switch on, and Gemini’s search grounding is the same[5][6]. So this check shows what the model says from what it has learned, which is often, not always, close to what a buyer sees in the app.
How to read your result
Named first
The answer led with you. Good, and worth checking twice: ask again and see whether you stay first. A brand that is named first every time is in a strong place. One that is first only sometimes is in rotation.
Named, but lower
Read the names above yours and the sentence that describes you. If the answer puts you in a narrower slot (“good for freelancers” when you sell to agencies), that wording usually comes from the pages the model learned from, and it is the description to change.
Not named, but rivals are
The most useful result. The card shows who was named instead, and the summary shows the name most often ahead of you across both engines. Those are the brands whose sources, reviews and comparisons to study first.
No brands named at all
The question may be too general or too informational for the engine to recommend anyone. Rephrase it the way a buyer would when ready to choose: add who it is for, a budget or a must-have feature.
What one run can and cannot tell you
If the list rotates, one check is a coin toss reported as a finding. Move the sliders to see how wide the honest range is at different numbers of runs, and when a gap between you and a rival stops being noise.
This is the reason the button under each result says “Ask again”. Run the same question three or four times and you will usually see the list change. That is not the tool misbehaving. It is what the engines do.
Which questions to check
The question matters more than anything else on this page. Check the ones a buyer asks when they are close to choosing, in their words, not yours. Your ticks are saved in this browser only.
For more on how ChatGPT chooses which products to name, see how ChatGPT product recommendations work, and for the wider practice, the ChatGPT SEO guide.
How to get named more often
There is no setting that puts you in an answer. What helps is being easy to find and easy to describe in the places models learn from and retrieve:
- Be in the comparisons. Lists, reviews and “best X for Y” articles are where an answer gets its shortlist. Citation analysis shows which sources the engines lean on in your category.
- Be mentioned, not only linked. A brand named in many credible places is easier for a model to recall. See brand mentions vs backlinks.
- Say plainly who you are for. If your own pages describe you vaguely, so will the answer. Clear category, audience and use cases on your site give the model the sentence to repeat.
- Let the crawlers in. A model cannot cite a page its crawler was refused. Check with the AI crawler checker.
From one check to tracking
This page answers “where was I in one answer today”. Tracking answers “how often am I named, and is it changing”. The same questions are asked every week, and after 4 weeks each brand is sorted into the ones named nearly every time and the ones that rotate in and out. That split is Core vs Carousel.
Every limit is on the pricing page. To see how share of voice and rivals are reported once tracking runs, read about AI visibility tracking and competitor tracking.
This tool or the AI visibility checker?
Use this one when you know the question. Use the AI visibility checker when you do not: it reads your site, writes five buyer questions for your category and runs them all, which gives a wider first look. Both ask the same two engines, and both show you the real answers behind every number.
Questions people ask
What is a ChatGPT rank tracker?
A tool that asks ChatGPT the questions your buyers ask and records whether your brand is named, where it appears among the names given, and who is named instead. Some call it an AI rank tracker or an LLM rank tracker. Unlike a Google rank tracker, there is no fixed list of results to read: the answer is written fresh each time.
Is there such a thing as a ChatGPT ranking?
Not a stable one. In SparkToro and Gumshoe's January 2026 study there was less than a 1 in 100 chance that ChatGPT gave the same list of brands in any two responses to the same prompt. What you can measure reliably is how often you are named across many runs, which is why this tool calls its number a position in one answer, not a rank.
Which models does this free checker use?
ChatGPT's budget API model (openai/gpt-5.6-luna) and Gemini's (google/gemini-2.5-flash-lite), each asked your question once with no system prompt and no web search. The ChatGPT app can search the web, so its answer to the same question can differ. The result card names the exact model that answered.
Why does ChatGPT name my competitor and not me?
Usually because the competitor appears more often in the sources a model learned from or retrieves: review sites, comparison articles, forums, directories and its own clear pages. Read the full answer for how it describes the brands it names. That wording is a good guide to what it has read about them.
Why did I get a different answer when I asked again?
Because answers are generated, not looked up. The same question can return a different list and a different order each time, which is the main finding of the SparkToro study. Ask again with the button under the result to see it happen.
How do I track my ChatGPT visibility over time?
Run the same set of questions on a schedule and count how often you are named. CoreCited's free plan does this for 5 questions a week on Claude and Gemini, and after 4 weeks it shows whether you are named every time (Core) or only sometimes (Carousel). Paid plans add ChatGPT and more AI surfaces.
Do you store my question or brand?
No. The question, brand, domain and competitors are sent to the two engines and the parse model, the result is returned to your browser, and nothing about the check is kept except a counter used for the hourly and daily limits.