AI Visibility Tracker: ChatGPT Search & LLM Brand Mentions

Find out whether AI assistants recommend your brand. For every prompt your customers ask, this Actor checks ChatGPT search on OpenAI's GPT-6 (live web, with citations) and what Grok 4.7 and DeepSeek V4 Pro recommend from their own knowledge (GLM-5.3 optional). Perplexity, Gemini and Google AI Overviews are optional and run with your own API keys (BYOK). It tells you:

• ✅ Is your brand mentioned? And at which rank compared to competitors
• 🔗 Is your website cited as a source? At which citation position
• 🥊 Which competitors are mentioned and cited instead
• 📊 Share of voice per engine across all prompts (summary report)
• 🙂 Sentiment of how the answer talks about you (optional)
• 🌍 Country and language targeting

Use it for GEO / AEO (generative engine optimization), brand monitoring, competitor research and client reporting.

Use cases

• SEO / GEO agencies: monthly client reports on how often ChatGPT recommends the client compared with competitors.
• SaaS and e-commerce brands: see which "best X" prompts mention you, and which sites ChatGPT cites instead, so you know which pages to win.
• PR and content teams: check whether launches, reviews and articles change what AI assistants say about you.
• Competitor research: share of voice for your whole category across engines.

Quick start

1. Click Start with the default input (Notion vs Coda and Obsidian). You get a full result in about a minute for about $0.10.
2. Replace brandName, brandDomain, prompts and competitors with your own.
3. Schedule it weekly and compare the SUMMARY reports over time.

Why this Actor

	This Actor	Single-engine scrapers
Engines per run	4 built in (ChatGPT search on GPT-6, Grok 4.7, DeepSeek V4 Pro, GLM-5.3) + 3 with your own keys (Perplexity, Gemini, AI Overviews)	1
Models	Current flagship models, not older/cheaper ones	Varies
Brand & competitor analysis	Built in (mention rank, citations, share of voice)	Raw answer only
Price per ChatGPT search check	$0.03	$0.09 or more per keyword on comparable Store Actors (Oct 2026)
Bring your own API keys	Yes: $0.005 per check	No
Reliability	APIs with live web search, no browser automation	UI scraping

Two kinds of visibility

• AI search visibility (chatgpt, perplexity, gemini, google_aio): the engine searches the web live and cites sources. Shows whether you are recommended and whether your site is used as a source.
• Model knowledge (grok_knowledge, deepseek_knowledge, glm_knowledge): the model answers from what it learned in training, without web search. Shows whether the model already "knows" your brand. URLs it writes are stored in mentionedUrls (from memory, may be outdated) and never counted as citations.

How it works

ChatGPT search uses the OpenAI Responses API with the web search tool. Perplexity (Sonar) and Gemini (grounding with Google Search) use their official APIs with your keys, and Google AI Overviews come from SerpApi with your key. Each answer is analysed for your brand, its aliases and your domain. A ChatGPT search check takes about 25–30 seconds (live web search); knowledge checks take 10–30 seconds. Each engine has automatic fallback to a sibling model if the primary one is unavailable, and the model field shows which model answered.

Important: answers from the official APIs are very close to, but not identical to, what a logged-in user sees in the ChatGPT, Perplexity or Gemini apps. AI answers also change between runs. Use repeats to measure stability, and read the results as a visibility indicator and trend, not as a screenshot of one user's screen.

Input

Field	Description
brandName	Your brand, e.g. Notion
brandDomain	Your main domain, e.g. notion.com (used to detect citations; subdomains count)
brandDomains	Other domains you own, e.g. notion.site for hosted docs or help centers
brandAliases	Other names: product names, abbreviations
prompts	Questions your customers ask, e.g. best note taking app for teams
competitors	[{"name": "Coda", "domain": "coda.io", "domains": ["coda.help"]}]
engines	chatgpt, grok_knowledge, deepseek_knowledge (default), glm_knowledge, plus perplexity, gemini, google_aio with your own keys
repeats	1–5 checks per prompt to measure answer stability
country, language	e.g. us, en
sentiment	Classify answer sentiment toward your brand
openaiApiKey	Optional: run ChatGPT search on your own OpenAI account at the BYOK price
perplexityApiKey, geminiApiKey, serpapiApiKey	Required for the Perplexity, Gemini and AI Overviews engines (BYOK price)

{
  "brandName": "Notion",
  "brandDomain": "notion.com",
  "brandDomains": ["notion.site"],
  "prompts": ["best note taking app for teams", "notion alternatives"],
  "competitors": [{ "name": "Coda", "domain": "coda.io" }, { "name": "Obsidian", "domain": "obsidian.md" }],
  "engines": ["chatgpt", "grok_knowledge", "deepseek_knowledge"]
}

Output

One dataset item per prompt × engine × repeat:

{
  "prompt": "best note taking app for teams",
  "engine": "chatgpt",
  "model": "gpt-6-luna",
  "webSearch": true,
  "answerPresent": true,
  "answer": "The best team note-taking app in 2026 depends on whether you need a flexible workspace… **Notion** …",
  "citations": [{ "url": "https://www.notion.com/help/guides/team-wiki", "domain": "notion.com" }, { "url": "https://help.coda.io/…", "domain": "help.coda.io" }],
  "mentionedUrls": [],
  "brandMentioned": true,
  "mentionRank": 1,
  "brandDomainCited": true,
  "brandCitationPosition": 1,
  "competitorsMentioned": ["Coda", "Confluence"],
  "competitorsCited": ["Coda"],
  "sentiment": "positive",
  "error": null
}

A summary report (key-value store record SUMMARY) gives, per engine and overall: mention rate, citation rate, share of voice and average mention rank for your brand and each competitor, plus the most cited domains.

Pricing (pay per event)

Event	Price
ChatGPT search check	$0.03
Model knowledge check (Grok 4.7, DeepSeek V4 Pro or GLM-5.3)	$0.01
Any check with your own API key (BYOK: Perplexity, Gemini, AI Overviews, or ChatGPT with your OpenAI key)	$0.005
Sentiment classification (optional)	$0.002

Failed checks are never charged. A Google query that shows no AI Overview is still charged, because the search was performed (answerPresent: false). Set a maximum cost per run in the run options and the Actor stops cleanly before exceeding it.

Typical runs:

Run	Checks	Cost
Default input (2 prompts × ChatGPT + Grok + DeepSeek)	6	≈ $0.10
20 prompts × (ChatGPT search + Grok + DeepSeek)	60	≈ $1.00
50 prompts, ChatGPT search only, weekly for a month	200	≈ $6.00

FAQ

Is this scraping ChatGPT? No. It uses APIs with live web search, so runs don't break when a website changes its UI.

Why do results differ from what I see in ChatGPT? Personalisation, memory, account location and model versions differ between the consumer apps and the API. Track trends over time and use repeats.

Can I schedule it? Yes. Use Apify Schedules to run it weekly and compare the SUMMARY reports over time.

Can AI agents call it? Yes. This Actor is enabled for x402 agentic payments: an AI agent can run it and pay per event in USDC on Base through the Apify API or MCP server, with no Apify account or API key. See Apify x402 docs.

More Actors by the same developer

• Government Tenders & AI Bid Matcher: open tenders from EU TED, UK, SAM.gov, Canada, Australia, World Bank and UN, scored by AI for your company.
• Google Trends API & Scraper + AI Trend Insights: interest over time, regions, related queries and trending searches with AI summaries.