Multi-AI Answer Snapshot: OpenAI, Perplexity, Gemini & Claude

Ask the same question to several AI assistants at once and see their answers side by side: what each one says, which websites it cites, and whether it mentions your brand or your competitors. Give this Actor a list of prompts (the questions your buyers ask AI) and your brand names. For every prompt you get one row with each assistant's answer, its cited sources, and a per-brand view of who mentioned you, in what order, and who cited your site.

Which assistants, and how (please read)

Every assistant is asked through its official API, with its own live web search:

Assistant	Model	Search	Model cost
OpenAI (the models behind ChatGPT)	gpt-5.4-nano	OpenAI web search tool	included in the price
Perplexity	Sonar	Perplexity's own search	billed to your Apify account, about $0.005 an answer
Google Gemini	Gemini 3.1 Flash Lite	Google Search grounding	billed to your Apify account, about $0.03 to $0.04 an answer
Anthropic Claude	Claude Haiku 5.5	Anthropic's web search	billed to your Apify account, about $0.013 to $0.025 an answer
xAI Grok	Grok 4.3	xAI's web search	billed to your Apify account, about $0.04 to $0.05 an answer

Perplexity, Gemini, Claude and Grok run through Apify's own OpenRouter proxy (apify/openrouter), so you need no API keys: Apify bills their model cost to your account at OpenRouter's rate. On Apify's free plan that cost is 10 times higher (Apify's rule for the proxy), so free-plan users should start with Perplexity and OpenAI. The run summary shows the model cost of every assistant.

These are the same models and live searches the consumer apps use, not copies of the apps: nothing is scraped from chatgpt.com, perplexity.ai or gemini.google.com, and an app's own answer to the same words can differ (it knows the user and their history). Answers also change between runs, which is why Monitor mode exists.

Not included: Microsoft Copilot and Meta AI have no public API for their assistant answers, so they cannot be asked legally and reliably. They are left out on purpose.

What you get for every prompt

• Each assistant's answer (openaiAnswer, perplexityAnswer, geminiAnswer, claudeAnswer, grokAnswer) and its status.
• Cited sources per assistant: link, site and title, plus allCitedDomains across all of them. Gemini's Google redirect links are resolved to the real site name.
• Your brands: for every name in Brands or websites to look for, per assistant: mentioned, rank (order of first mention among your names, so "Asana before Trello"), number of mentions, and cited (its website is one of the sources). mentionedBy counts how many assistants mention each name; brandVisibility lists which ones.
• Run summary: mention rate per brand per assistant, average rank, how often each brand's site is cited, the most cited websites overall (where to get listed or reviewed), and model cost per assistant.

Mentions are found in code by exact whole-word matching ("Shark" does not match "Sharks"), never by asking a model to judge.

Who uses this

• Brands and marketing teams checking whether AI assistants recommend them, or their competitors, for the questions their buyers ask.
• SEO and GEO agencies reporting AI visibility across assistants to clients, with a monitor that flags changes.
• PR and content teams finding which websites AI assistants cite most, to know where coverage counts.

Input

Give prompts as a plain list, an Apify dataset, a CSV or Google Sheet link (shared as "anyone with the link can view"), or inline rows. The prompt column is detected automatically (prompt, query, question, keyword...), or set Prompt field.

• AI assistants: any of openai, perplexity, gemini, claude, grok. Default: OpenAI, Perplexity and Claude.
• Brands or websites to look for: brand names (Asana) or websites (monday.com). A Brand field column can add names per row.
• User country: optional two-letter code so answers and sources fit that country.

Example (the default input):

{
  "prompts": ["What is the best project management software for a small team?", "Best cordless vacuum for pet hair"],
  "engines": ["openai", "perplexity", "claude"],
  "trackBrands": ["Asana", "Trello", "monday.com", "Dyson", "Shark"]
}

Output (one row per prompt, shortened)

{
  "prompt": "What is the best project management software for a small team?",
  "checkStatus": "ok",
  "summary": "3 of 3 assistant(s) answered. Mentioned: Asana by OpenAI, Perplexity, Claude; Trello by OpenAI, Perplexity, Claude; monday.com by OpenAI, Perplexity.",
  "enginesAnswered": ["openai", "perplexity", "claude"],
  "mentionedBy": { "Asana": 3, "Trello": 3, "monday.com": 2 },
  "openaiCitedDomains": ["zapier.com", "asana.com"],
  "perplexityBrandsMentioned": ["Asana", "Trello", "monday.com"],
  "answers": [
    { "engine": "perplexity", "model": "perplexity/sonar", "status": "ok", "citedDomains": ["pcmag.com", "asana.com", "zapier.com"],
      "brands": [{ "name": "Asana", "mentioned": true, "rank": 2, "mentions": 3, "cited": true }], "modelCostUsd": 0.00537, "modelCostPaidBy": "caller" }
  ]
}

Monitor mode: get told when it changes

Switch on Monitor mode and schedule the run (daily or weekly). The first run records each prompt's starting point. Later runs report, per assistant, what changed: your brand newly mentioned or dropped, moved in order, your site newly cited or no longer cited, and newly cited websites. Set a Webhook URL (Slack, Zapier, Make, n8n) and it is only called when something changed. If one assistant fails on a run, its last result is kept, so an outage is not reported as a change.

Pricing

Pay per event, no subscription:

• $0.03 per prompt answered (by any number of the chosen assistants). Lower on higher Apify plans. OpenAI's answer is included.
• $0.01 per CSV or Excel export file, only when you ask for one.
• $0.01 per webhook delivered (confirmed 2xx only; in monitor mode only when something changed).
• Plus the model cost of Perplexity, Gemini, Claude and Grok, billed by Apify's OpenRouter proxy (see the table above).

You are charged only when at least one assistant answers. Empty prompts, duplicates, prompts no assistant answered, and rows skipped because of a time or budget limit are free. Example: 100 prompts with the default three assistants on a paid plan is $3 here plus about $3 of Perplexity and Claude model cost.

Honest limits

• Answers come from each company's API with web search, not from their consumer apps, so they will not match one person's app exactly, and they vary between runs.
• Grok is the slowest (about 25 to 30 seconds an answer); the others take 4 to 12 seconds. Two prompts run at once by default.
• rank is the order your tracked names first appear in the text, not a position in a product list.
• Only your prompt text and country are sent to each provider.

FAQ

Is this the same as asking ChatGPT, Perplexity or Gemini in their apps? It uses the same companies' models and live searches through their official APIs, so it shows how those AI systems answer your prompts. It is not a screenshot of an app.

Why no Copilot or Meta AI? Neither offers a public API for its assistant answers. Getting them would mean scraping, which this Actor does not do.

Can an AI agent call it? Yes. Pay per event works through Apify's MCP server and x402, and the output is plain JSON.

Something looks wrong for one prompt. Open an issue on the Actor page with the prompt and I will look the same day.

If this saved you asking each AI assistant by hand, a review on the Store helps a lot.