AI Answer Engine Tracker: ChatGPT, Gemini and Claude answers with sources

Ask ChatGPT, Gemini and Claude the questions your customers ask, with web search on, and get each answer back as data: the full text, every source it cites, the searches it ran, and whether your brand, your website and your competitors show up.

• Official APIs, not scraping. Answers come from OpenAI, Google and Anthropic's own APIs with web search turned on (models: gpt-6-luna, gemini-3.5-flash-lite, claude-haiku-5-5). No browser automation, no account logins, no broken runs when a website changes.
• One record per prompt and engine, with the same fields for every engine.
• Brand visibility built in: named or not, how many times, how early in the answer, your domain cited or not, which competitors are named.
• Pay per answer. Failed answers are never charged.

Input

Field	Type	Description
prompts	array	Questions, one per line. Each goes to every engine you pick. Empty = a free format example.
engines	array	Any of chatgpt, gemini, claude (default: all three).
brand	string	Optional. Your brand name.
brandDomain	string	Optional. Your website, e.g. example.com.
competitors	array	Optional. Competitor names.

{
    "prompts": ["best project management tool for a small team", "asana alternatives"],
    "engines": ["chatgpt", "gemini", "claude"],
    "brand": "Asana",
    "brandDomain": "asana.com",
    "competitors": ["Trello", "ClickUp", "monday.com"]
}

Output (one record per prompt and engine)

{
    "prompt": "best project management tool for a small team",
    "engine": "claude",
    "model": "claude-haiku-5-5",
    "status": "ok",
    "answer": "There's no single best tool for every team, but three names come up again and again ...",
    "citations": [{"url": "https://project-management.com/...", "title": "...", "domain": "project-management.com"}],
    "citedDomains": ["plaky.com", "project-management.com", "wrike.com"],
    "searchQueries": ["best project management tool small teams 2026"],
    "brandMentioned": true,
    "brandMentionCount": 3,
    "brandFirstPosition": 0.04,
    "brandDomainCited": false,
    "competitorsMentioned": ["ClickUp", "monday.com"],
    "askedAt": "2026-10-20T10:56:12+00:00"
}

brandFirstPosition is where the first mention starts, as a share of the answer (0 = the first word).

Pricing

Pay per event, no subscription.

Event	Price	When it is charged
Answer	$0.045	Once per prompt and engine, after the answer arrived. 10 prompts x 3 engines = 30 answers = $1.35.

Failed answers are never charged. Set a maximum charge in the run options and the actor stops before it would go over. Apify's standard run-start event adds $0.00005 per run.

Limits, stated plainly

• API answers are not copies of the consumer apps. ChatGPT, Gemini and Claude in their apps add account memory, personalization and their own system prompts. This actor asks the same model families through their APIs with web search on. Answers are close in substance but can differ, and they vary from run to run, as the apps do.
• Every answer is based on a fresh web search: ChatGPT runs up to 2 searches, Gemini and Claude at least one.
• Gemini's citation links are Google redirect links (vertexaisearch.cloud.google.com/...); the domain field holds the real source site.
• Apify free plan: up to 3 answers per run (each answer costs us an API call). Paid plans have no cap.
• Prompts longer than a normal question are fine, but the actor is built for questions, not documents.

Privacy

Prompts are sent to the engines you pick (OpenAI, Google, Anthropic) and nowhere else. Answers are stored only in your run's dataset. No account data is read; the actor runs with limited permissions.