AI Visibility Tracker — ChatGPT, Gemini, Perplexity Brand Monitor

Does ChatGPT recommend your brand — or your competitor's? Track brand mentions, share of voice, and cited sources across AI answer engines. Built for GEO (Generative Engine Optimization), AI SEO, agencies, and brand managers.

No API keys needed for the keyless engines. Bring your own OpenAI or Google AI key to add ChatGPT and Gemini.

What it does

Send the prompts buyers ask AI ("best crm for small business") plus your brand and competitors. Every prompt × engine returns one structured row:

• mentionedBrands — which tracked brands appear in the answer, in order
• brandPositions — per-brand rank (1st mention = rank 1), character index, mention count
• primaryMentioned / primaryRank — did YOUR brand (first in list) get cited, and where
• citations / citedDomains — every source URL the engine cited (domain, title) — the list to pitch for PR/backlinks
• answerText — the full AI answer for manual review
• hasAiOverview — whether an AI answer rendered at all (absence is real signal)

Engines

Engine	Setup	Notes
duck_ai	None — keyless	DuckDuckGo's private AI chat (GPT-class model). Reliable default
google_ai_overview	None — keyless	Google's AI Overview SERP box. Best-effort; presence varies by query/geo
chatgpt	Your OpenAI API key	Web-search-enabled model — real citation URLs, exactly what ChatGPT cites
gemini	Your Google AI key	google_search grounding — free key from AI Studio works
perplexity	Optional	Official Sonar API with your key; anonymous endpoint tried without one (rate-limited)

One row per prompt per engine — a missing key produces an error row for that engine, never a failed run.

Example input

{
  "prompts": [
    "best crm for small business",
    "top project management tools 2026"
  ],
  "brands": ["HubSpot", "Zoho", "Pipedrive", "Asana", "Monday.com"],
  "engines": ["duck_ai", "google_ai_overview", "chatgpt", "gemini"],
  "openaiApiKey": "sk-...",
  "geminiApiKey": "AIza...",
  "proxyConfiguration": {
    "useApifyProxy": true,
    "apifyProxyGroups": ["RESIDENTIAL"]
  }
}

brands[0] is your brand — its mention/rank get dedicated fields. Keys are used for the run only and never stored.

Use cases

• GEO / AI SEO — is your brand in the answer when buyers ask AI? Track weekly
• Share of voice — mention rank vs competitors across every engine, in one dataset
• Source-gap analysis — citedDomains = the publications driving AI visibility; pitch them
• PR & launch monitoring — did AI answers change after your campaign?
• Agency reporting — scheduled runs export to CSV/JSON for client dashboards

Why this actor

Dedicated GEO platforms charge $100-500/mo and other Apify actors charge up to $0.08/result or require API keys for every engine. This tracker is $0.001/row and runs keyless out of the box.

Scheduling

Pair with an Apify Schedule (weekly) and your dataset becomes a time series of AI visibility — the metric GEO SaaS tools sell at enterprise prices.

Notes

• Residential proxies recommended for the browser engines — Google and anonymous Perplexity gate datacenter IPs
• AI answers are non-deterministic; judge trends over multiple runs, not single results
• Per-engine failures are recorded per row (error field) — one engine failing never kills the run

Example output

{
  "type": "ai_answer",
  "engine": "google_ai_overview",
  "prompt": "best crm for small business",
  "trackedBrands": [
    "HubSpot",
    "Zoho"
  ],
  "answerText": "",
  "answerLength": 0,
  "hasAiOverview": false,
  "citations": [],
  "citedDomains": [],
  "scrapedAt": "2026-10-03T23:34:03.265157+00:00",
  "mentionedBrands": [],
  "brandPositions": {
    "HubSpot": {
      "mentioned": false,
      "charIndex": -1,
      "count": 0,
      "rank": null
    },
    "Zoho": {
      "mentioned": false,
      "charIndex": -1,
      "count": 0,
      "rank": null
    }
  },
  "primaryMentioned": false,
  "primaryRank": null
}

Pricing

Pay per dataset row — $1 per 1,000 results. No subscriptions, no setup fees; you only pay for rows returned.

FAQ

Does AI Visibility Tracker need a login or API key? No — every row comes from public endpoints. Provide a proxy only if your network is heavily rate-limited.

How do I monitor on a schedule? Attach the actor to an Apify Schedule (daily/weekly) and use the watch/monitor mode if present — each run emits a fresh, diff-able dataset.

What does each row contain? See the output fields table above; every row carries a scrapedAt timestamp so datasets can be deduped across runs.

Is it legal to scrape this data? The actor only accesses publicly available data — the same pages/APIs a browser sees. Scraping public data for research, monitoring, and analysis is common practice, but platform Terms of Service and the rules in your jurisdiction still apply to how you use it. Consult legal counsel for your specific case.

More from this developer

Monitors & business intel: Otodom Property Monitor · Fotocasa Property Monitor · Subito.it Monitor · Kleinanzeigen Monitor · Copart Auction Monitor · AutoScout24 Car Monitor · Indeed Job Monitor · Upwork Job Monitor · ATS Jobs Monitor · Sports Scores & Odds Monitor · Prediction Markets Monitor · Crypto Token Monitor · Crypto Derivatives Monitor · Google News Monitor · App Store Monitor · Walmart Product Monitor · Shopify Competitor Monitor · Craigslist Listing Monitor · Airbnb Rental Monitor · Zillow Property Monitor · Pump.fun Monitor · StockTwits Scraper · Yahoo Finance Scraper · Product Hunt Monitor · Hacker News Monitor

Alt-social & OSINT: Gab Scraper · BitChute Scraper · Truth Social Scraper · Telegram Scraper · Discord Scraper · Rumble Video Scraper · Odysee Scraper · Kick Scraper · 4chan Scraper · Bitcointalk Scraper · Xiaohongshu / RedNote Scraper

Creator platforms: Bluesky Scraper · Threads Scraper · Farcaster Scraper · Spoutible Scraper · Pixelfed Scraper · Substack Notes Scraper · Dailymotion Scraper · Roblox Scraper

Gov & regulatory: Federal Register Monitor · Federal Court Monitor · FDA Compliance Monitor · Clinical Trials Monitor · Lobbying Monitor