AI Visibility Tracker: Brand Mentions in ChatGPT, Perplexity, Gemini

This AI visibility tracker shows what ChatGPT, Perplexity and Google Gemini answer when shoppers ask the questions that matter to your brand, and whether your brand makes the list. You give it your brand, your competitors and a few shopping prompts. It asks each assistant through its official API with your own keys and labels every answer: mentioned or not, recommended or not, rank among the products named, tone, cited URLs and the competitors that showed up instead. Run it on a schedule to see your AI share of voice change over time.

Who it's for

• Brand and ecommerce teams who want to know if AI assistants recommend them
• SEO and content people working on "AI search" or "generative engine" visibility
• Agencies that report brand share of voice across AI answers for clients
• Founders checking how their product is described next to competitors

What you get

One row per prompt and engine.

Field	What it means
prompt, engine, model	The question, the engine (chatgpt, perplexity, gemini) and the exact model that answered
answer	The full answer text
citedUrls, citedDomains	The sources the engine cited (ChatGPT and Perplexity; Gemini runs without web search and returns none)
brandSiteCited	True if a page from your own domain was cited
brandMentioned + probability	Does the answer mention your brand or its products
brandRecommended + probability	Does the answer recommend your brand as a good choice
brandRank	Position of your brand among the products the answer names (1 = first), empty if not listed
productsNamed, rankedProducts	All products and brands the answer offers, in order
sentiment + confidence	positive, neutral, mixed or negative tone toward your brand
competitorsMentioned	Which of your competitors appear, with their rank
error	Why a row has no answer (bad key, engine error). These rows are free

How to run it

1. Enter your brand, optionally its website domain and other product line names.
2. Add your competitors.
3. Write your own prompts, or pick templates like "Best {category} under {price}", "{brand} vs {competitor}" or "Alternatives to {brand}" and fill in category and price.
4. Add at least one API key: OpenAI, Perplexity or Gemini. Each key you add turns on that engine.
5. Set the country (2-letter code) so answers fit the market you sell in.
6. Start the run. A few prompts finish in well under a minute.

Why your own API keys?

The answers come straight from the official APIs, billed by OpenAI, Perplexity or Google to your account. That keeps the price per result low and means nothing is scraped from a chat website. Keys are secret inputs: Apify stores them encrypted and the Actor only uses them to send your prompts.

The default models are the providers' fast, low-cost ones. ChatGPT and Perplexity answer with web search, and those search calls are billed by the provider too, so check their pricing page for current rates.

Gemini is queried without Google Search grounding, so its answers come from the model's own knowledge and include no cited links. Google's terms for Grounding with Google Search do not allow storing or analysing grounded results, so this Actor never uses that feature. If you are in the EU, EEA, Switzerland or UK, Google requires a paid-tier Gemini API key.

Example output

Real result from a test run (answer and lists shortened):

{
  "prompt": "What are the best alternatives to Allbirds?",
  "engine": "chatgpt",
  "model": "gpt-5.4-mini-2026-03-17",
  "country": "US",
  "answer": "If you like Allbirds for their comfort, minimalist style, and eco-friendlier materials, the best alternatives depend on what you want to optimize for: ...",
  "citedDomains": ["wired.com", "goodhousekeeping.com", "gearpatrol.com"],
  "brandSiteCited": false,
  "brandMentioned": true,
  "brandMentionedProbability": 0.99,
  "brandRecommended": false,
  "brandRecommendedProbability": 0.14,
  "brandRank": null,
  "productsNamed": 11,
  "sentiment": "positive",
  "sentimentConfidence": 0.91,
  "competitorsMentioned": [
    {"name": "Hoka", "rank": 5, "probability": 0.99},
    {"name": "Brooks", "rank": 6, "probability": 0.99}
  ],
  "rankedProducts": [
    {"rank": 1, "name": "Everlane Court Sneaker / Tread", "isYourBrand": false},
    {"rank": 2, "name": "Veja", "isYourBrand": false},
    {"rank": 3, "name": "Adidas Stan Smith / Clean Court styles", "isYourBrand": false},
    {"rank": 4, "name": "Ecco", "isYourBrand": false}
  ]
}

How to use the AI visibility tracker via API

You can run the AI visibility tracker in the Apify Console, or call it from your own code through the Apify API. The API tab of this Actor has ready-to-copy examples for the JavaScript and Python API clients, the Apify CLI, plain HTTP and the OpenAPI definition. To use it from an AI assistant or agent, connect the Apify MCP server and add this Actor as a tool.

Send the same JSON input you would fill in the form, for example your brand, your competitors, your prompts and your API keys, start a run, and read the results from the run's default dataset as JSON, CSV or Excel. You can also schedule runs in Apify.

Pricing

You pay per prompt analyzed: one prompt answered by one engine and fully labeled. Three prompts on two engines are six results. Rows that fail (wrong key, engine error, empty answer) are not charged. The API usage itself is billed by the AI provider on your own key.

Tips

• Use the questions your customers really ask. "Best {category} under {price}" and "alternatives to {brand}" are good starting points.
• Run the same prompt set every week and compare brandMentioned and brandRank over time. AI answers change often, so one run is a snapshot, not a verdict.
• Add short product line names to "Other names for your brand" so a mention like "Tree Runner" counts.
• Short brand names (3 letters or less) are matched case sensitive, so "On" the brand is not confused with the word "on".
• brandRank counts distinct products in the order the answer lists them. For "{brand} vs {competitor}" prompts your brand is part of the question, so read brandRecommended and sentiment rather than rank.

FAQ

Does this scrape ChatGPT or other chat websites? No. It only uses the official APIs of OpenAI, Perplexity and Google. ChatGPT and Perplexity use their built-in web search. Gemini answers from model knowledge without web search.

Is it legal? The Actor sends your prompts to the providers' public APIs under your own API account, so their API terms apply to you as the key owner. It does not target personal data, although AI answers can name people such as founders. Please check the providers' terms for your use case.

Are the answers the same as in the ChatGPT or Gemini app? Not exactly. Apps can use different models, personalization and settings than the API. The API is the closest repeatable view, and it is the same for everyone who runs the same prompt. Gemini answers here come without web search, so they can differ more from the Gemini app.

How are mentions and recommendations decided? Exact name matches count first. On top of that an AI check reads each answer and returns a probability for each label, which is included in the output so you can set your own cut-off.

Which models are used? By default a fast, low-cost model per provider. You can choose another model in the Advanced section. For ChatGPT it must support the web search tool.

What if I only have one API key? That works. Only the engines you added a key for are queried.

How do I check if ChatGPT mentions my brand? Enter your brand, a few prompts your customers would ask, and your OpenAI API key. Each answer comes back with brandMentioned, brandRecommended and brandRank, each with a probability or position.

How do I track AI search visibility over time? Run the same prompt set on a schedule, for example every week, and compare brandMentioned and brandRank between runs. AI answers change often, so one run is a snapshot.

Disclaimer

This Actor is not affiliated with, endorsed or sponsored by OpenAI, Perplexity or Google. ChatGPT, Perplexity and Gemini are trademarks of their owners and are named only to describe which services the Actor queries with your own API keys.

About Reliable Tools

Reliable Tools builds Apify Actors that do one job well and keep working. See all our Actors at apify.com/reliable_tools.