apinoa
AI answer API · part of the Apinoa data API

What AI engines answer,
as JSON.

Ask ChatGPT, Copilot, Gemini, Google AI Mode, Google AI Overviews, Naver AI Briefing, and Naver AI Tab a question and get each answer back as data: its text, the sources it cites and the brands it names. $0.0012 per answer; failed calls are free.

POST/v1/chatgpt/answer {"prompt":"What are the best running shoes for beginners?","country":"US"}

    Updated 2026-11-08

    What can the AI answer API do?

    Send a prompt to any of 7 AI answer engines and get back what a person asking it would see, in one answer schema: the answer as Markdown and plain text, every source and whether it was cited, the passages that cite each source, and the brands and products the answer names.

    Ask from the market you care about with country, keep a conversation going with a follow-up, and compare engines field by field. It is the data behind AI visibility and GEO tracking: which brands an engine recommends, and which sites it trusts.

    Open the API reference

    Operations

    answer
    An engine's answer to a prompt, with its sources, citations and mentions.7 sources
    One request
    curl -X POST "https://gateway.apinoa.com/v1/chatgpt/answer" \
      -H "x-api-key: $APINOA_API_KEY" \
      -H "content-type: application/json" \
      -d '{"prompt":"What are the best running shoes for beginners?","country":"US"}'

    One answer, field by field

    Every engine answers in the same schema, so one parser reads ChatGPT, Copilot, Gemini, Google AI Mode, Google AI Overviews, Naver AI Briefing, and Naver AI Tab.

    What one answer returns

    markdown
    The answer as Markdown, and as plain text.
    sources
    Every source the answer drew on, and whether it was cited.
    citations
    The passages that cite a source, with the source each one points to.
    mentions
    The brands and products the answer names.
    suggestions
    The follow-up questions the engine offers.
    conversation
    An id to ask a follow-up in the same conversation.

    Engines and the countries they answer from

    Built for AI visibility data

    • 7 engines, one schema

      ChatGPT, Gemini, Copilot, Google's AI answers and Naver's answer in the same fields.

    • Sources and citations

      Which sites an answer drew on, which it cited, and the exact passage each citation backs.

    • Brands it names

      The brands and products in each answer, ready to count across prompts and engines.

    • Ask from their market

      Set country and the engine answers as it does for people there.

    • Follow-up questions

      Continue the same conversation, and read the follow-ups the engine suggests.

    • Pay for answers only

      $0.0012 per answer from a prepaid balance. A call that fails is not billed.

    Give your AI agent
    commerce, social and search data

    Add the MCP server to Claude Code, Cursor or any MCP client, or build on the TypeScript and Python SDKs. One key covers all of it, and every repository is public.

    • MCP server

      Give Claude, Cursor and other agents the API

      npx -y apinoa-mcp
      View on GitHub
    • TypeScript SDK

      Typed client for Node, Deno and Bun

      npm install @apinoa/sdk
      View on GitHub
    • Python SDK

      Sync and async client for Python

      pip install apinoa
      View on GitHub

    Pay for the requests you make

    No plans and no minimum. Top up from $20 and each request comes off your balance. The more you send in a day, the less each one costs.

    AI answers
    $0.0012 per answer

    Failed requests are free. Ask for $1 of free credit to evaluate, no card needed.

    What counts as a request, and the rest of the detail

    Price per request at each daily volume. The count resets at 00:00 UTC.
    12,000
    That day costs
    $12.00
    Per request, on average
    $0.001
    30 days at this pace
    $360

    Top up more, pay less

    Top-ups from $5,000 carry 8–20% in bonus credit, added the moment the payment lands.

    The same heavy workload every day?

    We can fix a monthly price for it. Talk to us

    Ready to build?

    Create a key and make your first call. Ask for free credit to evaluate — no card needed.

    Using an AI agent? Connect over MCP

    Frequently asked questions

    Which AI engines are covered?

    ChatGPT, Copilot, Gemini, Google AI Mode, Google AI Overviews, Naver AI Briefing, and Naver AI Tab. Each engine's page in the API reference lists its countries and options.

    What does one answer contain?

    The answer text as Markdown and plain text, its blocks in order, every source with whether it was cited, the citation passages, the brands and products named, and the engine's follow-up suggestions.

    Can I use it to track AI visibility (GEO)?

    Yes. Run the same prompts on a schedule and count the brands in mentions and the sites in sources. Every engine returns the same fields, so the comparison needs one parser.

    Can I ask as if from another country?

    Yes, with country. Each engine answers from the countries its page lists.

    How much does an answer cost?

    $0.0012 per answer at the base rate, less as your daily volume grows. There is no plan or monthly fee: you top up a balance and each answer is charged from it.

    Am I charged when a call fails?

    No. A call that fails is not billed.