ChatGPT Brand Tracker — ChatGPT SEO & AI Visibility Tool

Ask ChatGPT what your buyers ask, and find out whether it recommends you, ignores you, or names a competitor instead.

Buyers now ask ChatGPT "what's the best X for Y" before they ever open Google. Whatever ChatGPT answers is your shortlist position. This Actor runs any list of prompts and reports, per prompt: whether your brand was named, how many times, how early, exactly how it was described, and how that compares with a competitor set.

What does ChatGPT Brand Tracker do?

It sends each of your prompts to ChatGPT live, reads the full answer, and measures your brand in it:

• Mentions — whether your brand (or any alias) is named, and how often, matched on whole words, case-insensitively.
• Position — how early in the answer you first appear.
• Context — the passage around your first mention, i.e. the words ChatGPT uses about you.
• Share of voice — your mentions against the competitors you list.
• Citations (web-grounded mode) — every source ChatGPT cites, whether your domain is among them, and which competitor domains are.

Mentions are counted in ChatGPT's own prose. In web-grounded mode the answer also contains inline citation links; a brand name inside a link title or URL is a source, which is reported through brand_cited, competitors_cited and cited_domains rather than counted as a mention.

What you get per prompt

Field	What it tells you
brand_mentioned + brand_mention_count	Whether ChatGPT named you, and how often
brand_position_score	How early you appear, bucketed 1–10 (1 = first tenth of the answer)
brand_context	The passage around your first mention — the words used about you
share_of_voice_pct	Your mentions as a share of you + all named competitors (null when nobody is named)
competitor_mentions	Per-competitor mention counts in the same answer (competitors not named are omitted)
answer_text	ChatGPT's complete answer (turn off with includeAnswerText)
citations / cited_domains / citation_count	Cited sources — web-grounded mode only
brand_cited / brand_cited_urls	Whether one of your domains was cited, and which URLs
competitors_cited	Competitor domains that were cited
mode, answered, error, checked_at	Which mode ran, whether an answer came back, any error, and when

The run summary (OUTPUT in the key-value store) adds your overall mention rate, citation rate and, in grounded mode, a ranked table of the domains ChatGPT cites most (top_cited_domains).

Use cases

• AI shortlist monitoring — track weekly whether ChatGPT still names you for your core buying questions.
• Share of voice — quantify your presence against named competitors, prompt by prompt.
• Messaging audit — brand_context shows the words ChatGPT uses to describe you; correct them at the source.
• Category discovery — find which brands ChatGPT thinks lead your category.
• Digital PR targeting — in grounded mode, the cited-domains table is a ranked outreach list.
• Client reporting — dated, exportable AI-visibility evidence for retainers.

How to use ChatGPT Brand Tracker

1. Enter your prompts — write them the way a buyer would ask ("What is the best email marketing platform for a small business?").
2. Enter your brand name and any aliases (product names, legal entity).
3. Add competitor names to get share of voice.
4. Leave Web-grounded mode off for the standard, cheaper check — or turn it on and add your domains and competitor domains when you need the citation list.
5. Click Start and export the results as JSON, CSV or Excel, or schedule the run weekly.

Example input

{
  "prompts": [
    "What is the best email marketing platform for a small business?",
    "Which email marketing tool should a Shopify store use?"
  ],
  "brandName": "Mailchimp",
  "brandAliases": ["Intuit Mailchimp"],
  "competitorNames": ["Klaviyo", "Brevo", "Constant Contact"],
  "webSearch": false
}

Input configuration

Field	Type	Default	Description
prompts	array of strings	— (required)	Questions to ask ChatGPT, up to 200 per run (duplicates removed).
brandName	string	—	Brand to look for in each answer. Optional — leave it empty to just collect answers.
brandAliases	array of strings	[]	Other spellings or product names that count as a mention.
brandDomains	array of strings	[]	Your domains, e.g. hubspot.com. Used only in web-grounded mode (subdomains match).
competitorNames	array of strings	[]	Competitor brands to count; drives share_of_voice_pct.
competitorDomains	array of strings	[]	Competitor domains. Used only in web-grounded mode.
webSearch	boolean	false	Web-grounded mode: ChatGPT searches the live web first and returns cited sources. Billed as a separate, higher-priced event.
includeAnswerText	boolean	true	Include the full answer on every row. Does not change the price.

Example output

A real row (trained-knowledge mode; brand_context and answer_text shortened, line breaks removed):

{
  "prompt": "Which email marketing tool should a Shopify store use?",
  "answered": true,
  "mode": "trained_knowledge",
  "brand_mentioned": true,
  "brand_mention_count": 2,
  "brand_position_score": 5,
  "brand_context": "...easy integration with Shopify product catalogs. 3. **Mailchimp** (via third-party integration) - **Why?** Mailchimp is a widely used email marketing platform, but it do...",
  "competitor_mentions": [{ "name": "Klaviyo", "count": 2 }],
  "share_of_voice_pct": 50,
  "brand_cited": false,
  "brand_cited_urls": [],
  "competitors_cited": [],
  "citation_count": 0,
  "cited_domains": [],
  "citations": [],
  "answer_text": "For a Shopify store, some of the best email marketing tools to consider are:\n\n1. **Klaviyo** ...",
  "checked_at": "2026-09-23T03:05:37.100Z",
  "error": null
}

In web-grounded mode the citation fields are filled, for example:

{
  "brand_cited": true,
  "brand_cited_urls": ["https://forum.asana.com/t/how-to-show-task-status-in-progress-report/98361"],
  "competitors_cited": ["clickup.com", "trello.com"],
  "citation_count": 6,
  "cited_domains": ["guideflow.com", "clickup.com", "trello.com", "typesofsoftware.com", "forum.asana.com", "instagantt.com"],
  "citations": [
    { "title": "Best Free Project Management Software (2026)", "url": "https://clickup.com/learn/topic/project-management/tools/free/", "domain": "clickup.com" }
  ]
}

How much does it cost to track your brand in ChatGPT?

Two modes, two prices — pick deliberately.

Mode	What it measures	Citations	Price (Free & Bronze plans)
Trained knowledge (default)	What ChatGPT says from its own model — what most users actually get	No	$0.005 / prompt
Web-grounded	ChatGPT searches the live web first, then answers	Yes	$0.06 / prompt

Event	Price (Free & Bronze plans)
Actor start	$0.001 per run
Prompt check (trained knowledge)	$0.005 per prompt
Prompt check (web-grounded)	$0.06 per prompt

Apify Silver plans get 10% off and Gold plans 20% off these prices. A run is entirely one mode or the other, so only one of the two prompt events is ever charged in a run. A prompt that fails and returns no answer comes back with error set and is not charged. If every prompt in a run fails, the run ends as FAILED and nothing is charged, not even the start event.

Web search costs roughly 8.5× more to run and its cost is volatile — a citation-heavy prompt can cost several times a plain one. Rather than blend that into one inflated price, the two modes bill under separate events so an ungrounded run stays genuinely cheap.

Minimum-charge floor. If the prompts in a run turn out unusually expensive to answer — in practice this happens only in web-grounded mode — the Actor adds a small number of extra prompt-check units so the run is not sold below cost. The log states it and the OUTPUT record shows it as minimum_charge_units_added.

Worked examples:

Run	Charged
4 prompts, trained knowledge	$0.001 + 4 × $0.005 = $0.021
25 prompts weekly, trained knowledge	$0.001 + 25 × $0.005 = $0.126
25 prompts, web-grounded	$0.001 + 25 × $0.06 = $1.501 (plus any floor units)
3 prompts, web-grounded (real test run, floor added 1 unit)	$0.001 + 4 × $0.06 = $0.241

Most brand tracking wants the default. Turn grounding on when you specifically need the source list.

FAQ

Do I need an OpenAI API key? No. Everything is handled for you — add prompts and run.

Are these real ChatGPT answers? Yes, generated live per prompt. Nothing is simulated or replayed from a cache.

Why no citations in the default mode? Because ChatGPT is not searching the web — it is answering from trained knowledge, so there is nothing to cite. That is also what most users see. Enable web-grounded mode if you need sources.

Why do answers change between runs? LLM output is non-deterministic. Track the trend across weekly runs rather than reading one run as fact.

Do I have to supply a brand? No. Leave the brand fields empty and you still get every full answer.

Want other engines? See our Perplexity Brand Tracker, Google AI Overview Tracker and Google AI Mode Tracker.

Track every AI surface

Each Actor in this family covers one AI surface deliberately, so you only pay for the check you need:

• Google AI Overview Tracker — does Google's AI Overview cite you for your keywords?
• Google AI Mode Tracker — Google's conversational AI tab, tracked the same way.
• Perplexity Brand Tracker — full live answers plus the numbered citation list Perplexity is known for.
• AI Citation Gap Finder — which domains Google's AI cites across your whole keyword set, and where competitors are cited but you are not.
• AI Rank Tracker Pro — every platform in one run, with share-of-voice reports across ChatGPT, Perplexity, Gemini and Google AI.

---

Tags: chatgpt seo, chatgpt visibility, chatgpt api, chatgpt monitoring, llm seo, ai seo tool, aeo tool, ai search optimization, ai brand monitoring, ai brand visibility, ai visibility, brand visibility, GEO, AEO, generative engine optimization, answer engine optimization, share of voice, llm monitoring, ai search, ai mentions