How to Check If ChatGPT Mentions Your Brand (5 Free Ways)

To check if ChatGPT mentions your brand, ask it the questions your customers ask, in a temporary chat with personalization turned off, and record whether you appear, where, and how you're described. Then check which pages ChatGPT cites, confirm OpenAI's search crawler can reach your site, and look for visits tagged utm_source=chatgpt.com in your analytics. All five methods below are free, and the scoring rubric and tracking template turn one-off checks into a trend you can act on.

The steps reflect OpenAI's documentation as of October 2026. ChatGPT's interface changes often, so menu names may shift.

Quick checklist: 5 free ways to check ChatGPT brand visibility

1. Ask buyer questions in an unpersonalized temporary chat, several times each.
2. Review the cited sources when ChatGPT searches the web.
3. Confirm crawler access for OAI-SearchBot in robots.txt and at your firewall or CDN.
4. Find ChatGPT referrals in your analytics by filtering for chatgpt.com.
5. Use a free trial or free tier of an AI visibility tool to automate repeat checks.

Methods 1 and 2 show what ChatGPT says. Methods 3 and 4 show whether it can reach you and whether people click. Method 5 automates the repeat work.

Method 1: Ask ChatGPT the questions your buyers ask

This is the most direct check: you see the same kind of answer a prospective customer sees.

Write prompts the way customers talk

Asking "What is Acme?" only shows whether ChatGPT can describe you. The more useful test is whether it brings you up when someone who has never heard of you asks for help.

Write prompts across the types below. In these examples, Acme is a fictional project management tool for small agencies:

• Category: "What's the best project management software for a 10-person design agency?"
• Problem: "How can a small agency stop missing client deadlines?"
• Comparison: "Acme vs. [competitor]: which is better for agencies?"
• Alternatives: "What are good alternatives to [competitor]?"
• Local or niche: "Best project management tool for agencies in Chicago" (only if location matters to your buyers)

Category and problem prompts matter most, because the buyer hasn't picked a brand yet. Include the details real buyers add, such as team size, budget, industry, and must-have features. If you're stuck, our free prompt generator suggests buyer-style prompts for your category.

Run each prompt in a clean session

Your own ChatGPT account can skew results. To reduce that, use a temporary chat:

1. Open a new chat and select Temporary.
2. Choose Unpersonalized before you send your first message. OpenAI says this mode doesn't use your memories, custom instructions, or plugins. You can't change the setting after the conversation starts.
3. Paste one prompt and send it.
4. Start a new temporary chat for the next prompt, so earlier questions don't shape later answers.

OpenAI also says temporary chats don't show ads, so you won't confuse a sponsored unit with the answer itself.

Location matters too. OpenAI says ChatGPT may use an approximate location based on your IP address for local results, and that a VPN may affect the location it infers. If your market is regional, put the city in the prompt.

Ask each prompt more than once

Don't judge from a single answer. OpenAI's own developer documentation notes that generative AI is variable and that models sometimes produce different output from the same input. Run each prompt at least three times and score every answer with the rubric below.

Method 2: Check which sources ChatGPT cites

When ChatGPT searches the web, it shows where its information came from. Those citations tell you which pages feed answers in your category.

OpenAI says ChatGPT may search automatically when a question would benefit from current information. To make sure it searches:

1. In a conversation, select View all tools, then Search. You can also type / in the message box and choose Search.
2. Send your buyer prompt.
3. To rerun an existing answer with web results, use the refresh control and choose Search the web, when available.
4. Select a citation to open it, or select Sources to see the cited pages and other relevant links.

For each answer, note:

• Whether your own site is cited, and which page.
• Which third-party pages are cited: review sites, comparison articles, directories, forums, or news.
• Whether competitors' pages are cited when yours aren't.

If the same few pages keep appearing for your category, those are the pages worth getting into.

Open the sources instead of trusting the summary. OpenAI warns that search results and citations can be incomplete, outdated, or incorrect. If ChatGPT gets your pricing or features wrong, the cited page may show where the error came from.

Method 3: Make sure OpenAI's crawlers can reach your site

If OpenAI's search crawler is blocked, your pages won't be shown in ChatGPT search answers. Three OpenAI user agents matter here, and OpenAI's crawler documentation describes each one:

OpenAI treats these settings independently. You can allow OAI-SearchBot so you appear in search results while disallowing GPTBot to opt out of training. For showing up in ChatGPT search, OAI-SearchBot is the one to check first.

To check your setup:

1. Open yourdomain.com/robots.txt in a browser.
2. Look for groups that name OAI-SearchBot, GPTBot, or ChatGPT-User, and check whether any of them say Disallow: /.
3. Check the User-agent: * group too. Under the robots.txt standard, a crawler with no group of its own follows the * group, so a Disallow: / there also blocks OAI-SearchBot unless it has a group of its own.
4. If you want to appear in ChatGPT search, add an explicit rule:

User-agent: OAI-SearchBot
Allow: /

1. Check your firewall or CDN. OpenAI says to confirm that your host or content delivery network allows traffic from its published search bot IP addresses. A clean robots.txt won't help if bot protection blocks the requests.

After you change robots.txt, OpenAI says it can take about 24 hours for its search systems to adjust.

Our free AI visibility checker reads your robots.txt and reports the rules for GPTBot, OAI-SearchBot, ChatGPT-User, and other AI crawlers. It checks robots.txt rules, not firewall settings, so still do step 5 yourself. For background on the training crawler, see our GPTBot glossary entry.

Method 4: Find ChatGPT referral traffic in your analytics

Mentions that include a link can send you visitors. OpenAI's publisher FAQ says ChatGPT automatically adds utm_source=chatgpt.com to referral URLs from ChatGPT search results, so you can track those visits in tools like Google Analytics.

In Google Analytics 4:

1. Open the Traffic acquisition report under Acquisition.
2. View it by Session source or Session source/medium. Google's documentation says these dimensions show your utm_source values.
3. Find chatgpt.com in the list.

Then look at which landing pages those sessions start on. Those are pages ChatGPT linked to, which tells you what content it's citing. In other analytics tools or in server logs, search landing-page URLs for utm_source=chatgpt.com.

Keep two limits in mind:

• Analytics counts clicks, not mentions. If ChatGPT recommends you and the reader later searches your name or types your URL, that visit won't be credited to ChatGPT. Low ChatGPT traffic doesn't mean ChatGPT never mentions you.
• Keep paid and unpaid traffic apart. If you run ChatGPT ads, OpenAI says you can add your own tracking parameters, such as UTM tags, to ad landing-page URLs. Use distinct values so ad clicks don't blur your picture of unpaid referrals.

Method 5: Use a free trial or free tier for ongoing tracking

Manual checks give you a snapshot. Tracking dozens of prompts several times a month by hand gets slow, and that's the work AI visibility tools automate: they run a fixed set of prompts on a schedule and log mentions, positions, competitors, and cited sources over time.

Several tools offer free trials or free tiers, but terms and prices change often. Our AI visibility tools comparison lists each tool's pricing and the date we last verified it.

What to check before you rely on any ChatGPT visibility checker:

• Load the prompts you wrote for method 1, so you can compare the tool's results with what you saw yourself.
• Check which AI engines it covers beyond ChatGPT, if other engines matter to your buyers.
• Ask how it collects answers, for example through the ChatGPT interface or through OpenAI's API, and whether web search is turned on. That affects which sources show up.
• Look for citation data, not just a yes-or-no mention.
• Export your data before the trial ends.

How do you score ChatGPT's answers?

A yes-or-no mention hides most of what matters. Being named first with an accurate description is very different from a passing mention with last year's pricing. Score every answer on the same signals:

Score position from the answer's own structure. If ChatGPT gives a numbered list, the first item is a 3 and the rest are 2s. If it leads with one pick and lists alternatives, the pick is the 3.

Two numbers summarize a check. Your mention rate is answers that mention you divided by total answers. Your average position score is the mean of the position column across all answers, including zeros. Compare both with your top competitors to get your AI share of voice.

Treat accuracy separately. One major error repeated across many answers can matter more than a small change in mention rate, because it's what buyers read about you.

How many prompts should you test, and how often?

ChatGPT's answers vary, so every check is a sample, not a census. The goal is a sample big enough that a real change stands out from normal variation.

How many prompts and runs

Start with 20 to 30 prompts spread across the five prompt types, weighted toward category and problem prompts. Run each prompt three times per check. That gives 60 to 90 scored answers.

The arithmetic explains why. With 3 answers for one prompt, each answer moves that prompt's mention rate by a third, so a single prompt can swing from 33% to 67% on one different answer. With 60 answers across the set, one answer moves the overall rate by less than 2 points. Judge trends on the full set, and use individual prompts to find where to dig in.

If your category is small and the list of real buyer questions is short, test fewer prompts but run each one more times. More runs per prompt tighten each prompt's estimate.

How often to check

Monthly works for most brands. Check weekly when you're actively changing something, such as fixing robots.txt, publishing comparison content, or getting listed on a review site, and want to see when it lands. After a robots.txt change, allow the roughly 24 hours OpenAI mentions (see method 3) before rechecking.

How to compare over time

Hold everything constant between checks: same prompt wording, same temporary chat mode, same location settings, same number of runs. If you change the method, start a new baseline instead of comparing against the old one.

Treat small moves as noise. If your mention rate shifts by only one or two answers between checks, wait for the next check before reacting. Act when a change persists across two or more checks, or when a pattern shows up across a whole prompt type.

How to track brand mentions in ChatGPT week to week

A spreadsheet is enough. Copy these columns into Google Sheets or Excel, one row per answer:

Keep the prompt wording on a second tab keyed by Prompt ID, so it never drifts. Add a third tab called "Changes" where you log what you did and when: a robots.txt fix, a new comparison page, a review-site listing. That's how you connect a later shift in results to a specific action.

Each check:

1. Run every prompt three times in fresh, unpersonalized temporary chats, in Search mode if you're tracking citations.
2. Score each answer with the rubric and add one row per answer.
3. Calculate the mention rate and average position score for the whole set and for each prompt type.
4. Count competitor mentions and list the domains cited most often.
5. Note any accuracy errors and the cited page each error seems to come from.
6. Compare with the previous check, read the Changes tab, and pick one or two actions.

Worked example: checking Acme's ChatGPT visibility

Illustrative only. Acme is a fictional company, and every number below is made up to show the method. These are not real test results.

Acme sells project management software for small agencies. Its marketing lead writes 20 prompts and runs each three times in unpersonalized temporary chats with Search on, for 60 answers.

What the numbers say:

• ChatGPT knows Acme but rarely recommends it. Most mentions come from comparison prompts that already name Acme. In category and problem prompts, where buyers haven't picked a tool, Acme almost never appears.
• The same sources keep showing up. Category answers repeatedly cite two "best project management tools for agencies" roundups. Neither lists Acme.
• There's an accuracy problem. Two answers quote a price Acme retired last year, and the cited page is an old Acme blog post.
• robots.txt is fine. OAI-SearchBot is allowed, and the firewall allows OpenAI's published IP addresses.

Acme's plan: ask the two roundup publishers to review the product, update the old blog post with current pricing, and publish a page that answers the problem prompts directly. The team logs each action on the Changes tab and reruns the same 20 prompts monthly. Because category prompts produce 18 answers per check, they'll look for a change of several mentions that holds across two checks before calling it progress.

Common mistakes when testing ChatGPT visibility

1. Testing only your brand name. It shows ChatGPT can describe you, not that it recommends you.
2. Testing in your everyday account. A personalized chat can use your memories, and OpenAI says saved memories may shape the search queries ChatGPT writes. Use unpersonalized temporary chats.
3. Judging from one answer. Answers vary between runs. Run each prompt several times and look at rates.
4. Changing the setup between checks. New wording, a different mode, or a different location makes results hard to compare.
5. Counting an ad as a mention. OpenAI says ads can appear below a response for Free and Go users and are labeled as sponsored. Score only the response itself.
6. Ignoring the citations. The cited pages show where ChatGPT found its information, and they're often where the fix is.
7. Mixing up GPTBot and OAI-SearchBot. Blocking GPTBot opts you out of training, not search. Blocking OAI-SearchBot keeps you out of ChatGPT search answers.
8. Reading low referral traffic as "no mentions." Many mentions don't produce a tracked click.
9. Testing from one place when you sell in many. ChatGPT may use your approximate location for local results, so test the cities or regions you serve.

What to do if ChatGPT doesn't mention you

Start with what you control, then work outward:

1. Fix access. Allow OAI-SearchBot in robots.txt and make sure your firewall or CDN isn't blocking OpenAI's published IP addresses.
2. Study the cited sources. The pages ChatGPT cites for your prompts show where it finds information about your category. Work on getting included there, whether that's a review site, a comparison article, or an industry directory.
3. Answer buyer questions on your own site. Make it easy to see what you do, who it's for, what it costs, and how you compare with alternatives.
4. Correct errors at the source. If ChatGPT repeats an outdated fact, find the page it cites and ask for an update.

There's no shortcut. OpenAI says ChatGPT ranks search results using multiple factors meant to help people find relevant, reliable information, and that placement isn't guaranteed. For a fuller playbook, see our guide to ranking in ChatGPT.

Frequently asked questions

Why does ChatGPT give different answers each time?

OpenAI says generative AI is variable and that models sometimes produce different output from the same input. In ChatGPT, answers can also depend on whether it searched the web, your approximate location, and, if personalization is on, your saved memories, which OpenAI says ChatGPT may use when rewriting search queries. That's why you should test in unpersonalized temporary chats and repeat each prompt.

Is there a free ChatGPT visibility checker?

You can run every check in this guide for free: buyer prompts in temporary chats, citation review, a robots.txt check, and analytics filtering. Our free prompt generator helps you write the prompts, and our AI visibility checker checks your crawler rules. For automated tracking over time, several AI visibility tools offer free trials or free tiers.

Does ChatGPT use Bing or Google results?

OpenAI says ChatGPT search "sometimes partners with other search providers," sending them rewritten versions of your question. Its help article points to Microsoft's and Shopify's privacy policies for how those providers handle queries; as of October 2026, it doesn't mention Google. OpenAI also runs its own crawler, OAI-SearchBot, to surface websites in ChatGPT search, and says sites that block it won't be shown in ChatGPT search answers.

Can I pay to appear in ChatGPT's answers?

Not in the answers themselves. As of October 2026, OpenAI sells ads in ChatGPT, which can appear below a response for Free and Go users in select regions. OpenAI says ads are labeled as sponsored, are visually separated from the response, and don't influence ChatGPT's answers, and that advertisers can't shape, rank, or alter responses. Plus, Pro, Business, Enterprise, and Edu accounts don't see ads.

If I block GPTBot, will my site disappear from ChatGPT search?

Not according to OpenAI. GPTBot covers whether your content may be used for training, while OAI-SearchBot covers ChatGPT search. OpenAI says each setting is independent, so you can block GPTBot and still allow OAI-SearchBot.

How often should I check if ChatGPT mentions my brand?

Monthly is enough for most brands, using the same prompts and setup each time so results are comparable. Check weekly while you're making changes you want to measure, such as a robots.txt fix or new comparison content. Act on changes that hold across two or more checks, not on a single run.