aivisibilityskill API · for teams building AI visibility tools

Building an AI visibility tool? Skip the scrapers.

One API returns what 8 AI engines say for any question, parsed into the brands they name, their rank and sentiment, and every source they cite, with the raw response kept as a receipt. From about 1 cent an answer. You build the dashboard; we keep the engines working.

Get a free sandbox key Read the API reference

Sandbox keys return sample answers for free · 500 free credits for real ones · Volume from $500 a month: talk to us

Answers from

ChatGPT

Gemini

AI Overviews

AI Mode

Perplexity

Claude

Grok

DeepSeek

Request

curl https://api.aivisibilityskill.com/v1/checks \
  -H "Authorization: Bearer $AIV_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "questions": ["best invoicing app for freelancers"],
    "engines": ["chatgpt", "gemini",
                "ai_overviews", "ai_mode"],
    "mode": "live",
    "locale": "en-US",
    "wait": 60
  }'

Response, shortened200

{
  "status": "succeeded",
  "credits": { "quoted": 100, "charged": 100 },
  "answers": [{
    "engine": "chatgpt",
    "has_ai_answer": true,
    "brands": [
      { "name": "FreshBooks", "rank": 1,
        "recommended": true, "sentiment": "positive" }
    ],
    "citations": [
      { "position": 1, "domain": "nerdwallet.com",
        "source_type": "media" }
    ],
    "receipt": { "sha256": "9f2c…", "url": "…/receipt" }
  }]
}

Fig. 1One request asks four engines a buyer question. Field names are the API's own; the response is shortened. Full schema in the API reference.

Everything you'd otherwise build, and babysit.

Getting one ChatGPT answer is easy. Getting thousands from 8 engines every week, parsed the same way, retried when they fail and priced so you know your margin, is the part that eats your roadmap.

8 engines, one answer objectChatGPT, Gemini, Google AI Overviews, AI Mode, Perplexity, Claude, Grok and DeepSeek come from different providers in different formats. You get one schema for all of them.
Brands, ranks, sentiment and citationsEvery answer is read the same way: which brands it names, in what order, whether it recommends them, and every source it cites, with the source type.
Receipts your customers can checkEach answer keeps the engine's raw response and its SHA-256 hash, so any number in your dashboard traces back to what the engine actually said.
You never pay for failuresEngines time out and rate-limit. We retry, and anything that still fails isn't charged. You never pay for an error.
Cost control built inFree estimates before any spend, per-key daily budgets, confirmation thresholds and idempotency keys, so a bug in your code can't drain your balance.
Batch, standard or livePick speed per request: batch answers from 11 credits within a few hours, live answers at 25 credits, usually in under a minute.
Projects and analyticsTrack a brand, its competitors and its questions on a schedule, then read mention rates with error bars, competitor share, cited sources and history.
Free sandbox keysTest keys return realistic sample answers at no cost, so you can build and test your whole product before spending a credit.

Who builds on it

AI visibility dashboardsShow your customers where they're named across 8 engines without building or maintaining a single scraper.

SEO tools adding an AI tabPut AI answers next to your rankings data, with citations you can match against the pages you already track.

Agencies' client reportingRun your clients' buyer questions on a schedule and build your own reports, with receipts behind every number.

Research and data teamsCollect answers at batch prices to study how engines cite, rank and describe brands over time.

The endpoints

REST over HTTPS with Bearer keys, an OpenAPI 3.1 spec, a TypeScript SDK and an MCP server at https://mcp.aivisibilityskill.com/mcp. Prices are in credits; $1 buys 1,000 on a pack, and plans include more.

Main endpoints and what they cost
Endpoint	What it does	Credits
`POST /v1/estimate`	Price a check before running it	free
`POST /v1/checks`	Ask questions across engines	per answer
`GET /v1/jobs/{id}`	Answers with brands, ranks and citations	free
`GET /v1/answers/{id}/receipt`	The engine's raw response	free
`POST /v1/projects`	Track a brand, competitors and questions, with a schedule	free
`GET /v1/projects/{id}/visibility`	Mention rates per engine, with error bars	5
`GET /v1/projects/{id}/competitors`	Share of answers for each competitor	10
`GET /v1/projects/{id}/sources`	The sites engines cite for the project	5
`POST /v1/projects/{id}/site-check`	Read a site the way AI crawlers do	20
`POST /v1/projects/{id}/opportunities`	Rank what's keeping a brand out of answers	30
`POST /v1/opportunities/{id}/brief`	A fix plan an agent can apply	40

Credits per answer, by engine and speed
Engine	Batch	Standard	Live
ChatGPT, Gemini, Google AI Overviews, Google AI Mode	11	15	25
DeepSeek	–	–	25
Perplexity	–	–	35
Claude	–	–	60
Grok	–	–	200

Start free. Grow into volume.

Every account gets the same API. Bigger plans buy more credits per dollar and higher limits.

Start Free 500 credits and unlimited sandbox calls
- Every endpoint and engine
- 20 requests a minute
- No card
Get a key
Self-serve $9 to $299/mo 10,000 to 400,000 credits a month, or packs from $10
- 30 to 300 requests a minute
- Up to 1,000 tasks running at once
- Up to 2 years of history
See plans
Volume From $500/mo Committed monthly spend, for products built on our data
- Lower per-answer rates
- Higher rate limits and concurrency
- 99.5% uptime commitment with service credits
- Invoicing and a direct line to the founder
Talk to us

Built to be depended on

Public status page. Every part of the service is checked every 5 minutes, with 90 days of history and each engine's success rate: aivisibilityskill.com/status.
An uptime commitment. 99.5% monthly availability for the API and MCP server on volume plans, with service credits if we miss it: read it.
You only pay for answers. Failed work is never charged, and every paid call can be priced first for free.
Proof for every number. Each answer keeps the engine's raw response with a SHA-256 hash, so any number can be checked against its source.

Questions from builders

Can I show the results inside my own product?

Yes. You can use what the API returns, including answers, brands, citations and analytics, in your own product and reports, under your own brand. You can't resell raw access to the API itself or share your API keys with your customers. The details are in our Terms.

Which engines and countries do you cover?

ChatGPT, Gemini, Google AI Overviews, Google AI Mode, Perplexity, Claude, Grok and DeepSeek, in English for the United States, the United Kingdom, Canada, Australia and India.

How fast are answers?

Live answers usually arrive in under a minute, standard ones within minutes, and batch ones within a few hours. You choose per request, and the price follows the speed.

Do you have webhooks?

Not yet. Poll GET /v1/jobs/{id}; every response says when to check again in poll_after_seconds. Webhooks are on the roadmap, and volume customers can tell us what they need.

What happens when an engine fails or blocks a request?

We retry it. If it still fails, you aren't charged for that answer, and the job tells you which answers failed and why.

How do I get volume pricing?

From $500 a month of committed spend we offer lower per-answer rates, higher rate limits, an uptime commitment and invoicing. Email ahmed@aivisibilityskill.com with your expected monthly volume.

for your own product

Just want to check your own site?

Type your domain and see which products ChatGPT and Google's AI name for your buyers' questions, in about a minute.

Build on the AI answers. We'll run the engines.

Start with a free sandbox key and 500 credits. When your product grows, volume pricing and an uptime commitment are a reply away.

500 free credits, no card. Sign in with Google or email.