answers, grounded in what we publish
Frequently Asked Questions
- What is an AIR score?
AIR — Artificial Intelligence Ranking — is a domain's visibility on a 1-10 decile scale: a true NTILE(10) over how often a domain turns up in the AI's answers, not a score out of 100 on an invented formula. A domain needs at least 3 occurrences before it is scored at all. A score of 0 means insufficient data, not a bad result, and it is never coloured like one.
- How is the AIR score calculated, and where does the data come from?
We ask the AI real product- and service-related questions, run each phrase across multiple models, and store the complete answer as an immutable observation — the exact question, the full answer, every brand mentioned or ranked, and every domain and page cited. Phrase rankings, brand visibility, domain citations, and the AIR score are all derived from those stored observations. Nothing is estimated from SEO metrics or inferred; every number traces back to a stored answer.
- How often do rankings update?
Continuously, as observations accumulate — not on a fixed daily or weekly schedule. A ranking changes when new observations for that phrase or domain are added to the active dataset.
- Why does AIRANKS publish confidence intervals with its rankings?
Because a ranking is a measurement from a finite, variable sample of answers, not a fixed fact. Model responses vary run to run, so we resample the runs collected for each phrase 1,000 times — with replacement — recomputing the ranking fresh each time, and publish the resulting interval and stability percentage next to every rank. A wide interval or low stability means the ranking is close to a coin flip; a narrow interval and high stability means it held up across resampling. Below 25 runs we publish no interval at all rather than print a number the sample can't support.
- Is the data trustworthy? Is AIRANKS legit?
Every observation is stored and immutable — the exact answer, its timestamp, and every brand or domain it named — so a ranking can be checked against what was actually recorded. The methodology, the measurement principles in the Constitution, and the minimum-data thresholds (3 domain occurrences, 5 brand mentions, 25 observations before a phrase publishes stability) are all published, not hidden. A score of 0 is labelled insufficient data rather than folded into a bad score, and a ranking below the publish floor shows no interval at all rather than a false one.
- How much does AIRANKS cost?
We do not publish a price list here. What the site gives you today is the live ledger, phrase and brand rankings, per-domain reports, the methodology, the AIR Report, and a read API. The API and the AIR Report need an account so automated abuse stays out; the ledger, the rankings, and the AIR Toolbar need none. Anything that costs money will say so at the point you are asked for it.
- Does AIRANKS have a public API?
Yes — a read API for domains, brands, phrases, and search, documented at /docs/api with a machine-readable OpenAPI specification at /docs/api.json. It requires an account: mint an API key at airanks.net/tokens or sign in from the CLI with "air login". The AIR Toolbar needs no account. Responses carry ETag and cache-control headers. Official client libraries: the airanks Python SDK ("pip install airanks") and the @airanks-net/sdk JavaScript package on npm.
- Is there a command-line tool?
Yes — the open-source (MIT) airanks CLI. Run "npx airanks stripe.com" for any domain's AI Rank straight from your terminal, install it with "npm i -g airanks" to get the "air" command, or use Homebrew: "brew install airanks-net/tap/air". It looks up scores and searches phrases and brands — sign in with your account ("air login", or paste an API key from airanks.net/tokens) to use it; the same live data as the toolbar.
- What is the AIR Toolbar?
A browser extension that shows any site's AI visibility in one click — its AIR score, its AI-file audit, robots.txt crawler verdicts for GPTBot, OAI-SearchBot, ClaudeBot, and PerplexityBot, and the AI's own summary of the site, pulled live from the same public data behind every domain report. It reads only the current tab's hostname and talks to no origin other than airanks.net.
- Does AIRANKS support llms.txt?
Yes — we publish our own llms.txt, llms-full.txt, and ai.txt, and every domain report audits whether the site being reported on has them too, alongside robots.txt AI-agent verdicts, JSON-LD, and sitemap status.
- How do I improve my website's AIR score?
Start with your own domain report at airanks.net/d/yourdomain.com — it shows the AI's summary of your site, which pages have been cited, recent observations, and an AI-file audit covering llms.txt, robots.txt AI-agent access, JSON-LD, and sitemap status. Improving your AIR score means increasing how often the AI's answers mention your brand and cite your domain or specific pages, not traditional SEO metrics on their own.
- Who is behind AIRANKS?
AIRANKS was founded by Jeremy Schoemaker, who also builds the AI/LLM infrastructure behind the site — the agents, evaluation harnesses, and model plumbing that collect and score every observation.
- Is airanks.net the same company as airanks.com or airanks.ai?
No. AIRANKS operates only at airanks.net — we have no relationship with any other site using a similar name at a different domain, and our methodology, AIR score, and Toolbar belong to this domain alone. If you spot our work attributed to another domain, let us know.