Is the AI visibility scan really free?

Yes. The 18-check scan is free for any domain, with no account and no credit card. You get the full report, every fix, and a permanent shareable link. The email box on the home page sends a courtesy copy of the report; it is not a paywall.

What does Monitor at $199 per month include?

One domain. Every day, up to 50 of your prompts are asked on ChatGPT, Claude, Perplexity and Gemini (via their APIs with web retrieval), and each answer is checked for mentions of you and up to five named competitors, position, sentiment and citations, with a readable excerpt kept as evidence. Every rate carries its sample size and a Wilson 95% confidence interval. The 18-check re-scan runs every Monday with drift detection and an alert email on a major drop. AI-crawler traffic analytics come from access logs you upload.

What does Agency at $1,199 per month include?

Everything in Monitor across up to 25 client domains, with white-label reports, team seats and API keys. The prompt allowance is 250 active prompts pooled across the workspace: with all 25 clients enrolled that is an average of 10 prompts per client, against Monitor's 50 for one domain. You decide how to allocate them.

Yes, any time, from the Stripe billing portal (linked from Settings and your receipt) or by emailing support@aaivisibility.com. When Stripe ends the subscription, re-scans, AI-answer sampling and Monitor email stop and nothing further is charged; your saved reports and scan history stay readable. Prices exclude any tax Stripe applies at checkout.

Why does the free scan not show what percentage of AI answers mention my brand?

Because a single run is not a rate. AI answers vary run to run, so a percentage from one measurement session has no statistical meaning. Monitor samples repeatedly over time and reports every rate with its sample size and a 95% confidence interval; no headline number ships without its interval.

Which currencies and payment methods do you accept?

Prices are in US dollars. Payment is by card through Stripe Checkout; card details are entered on Stripe and never touch our servers. Promotion codes can be applied at checkout.

PRICING · SELF-SERVE · CANCEL ANYTIME

Track when AI answers mention your brand — and see the evidence.

The 18-check scan is free forever, for any domain. Monitor samples what ChatGPT, Claude, Perplexity and Gemini actually say about you every day, keeps an excerpt of every answer as evidence, reports every rate with its sample size and interval, and re-scans your site every Monday.

WHAT YOU RECEIVE

The first week on Monitor

Exact limits are in the plan table below; what is measured, engine by engine, is under “What is measured”.

Within a minute
Your brand name, aliases, competitors and first prompts are confirmed; the 18-check baseline scan of your homepage lands with every fix ranked.
Within minutes
The first AI-answer sample, started by checkout itself and retried by the 08:00 UTC daily run if it fails: every active prompt asked on four engines, each answer checked for mentions, position, sentiment and citations, with an excerpt you can read.
By day seven
A daily series with n and a 95% interval on every rate, the sources engines cite in your category, and the first Monday re-scan with drift detection.

Scan

No credit card. No account.

Full 18-check visibility report
Every fix, spelled out
Unlimited domains
Permanent, shareable result link

Run a free scan

Monitor

$199/mo

Per domain. Cancel anytime.

Weekly re-scans with drift alerts
Share of voice vs named competitors
Custom prompt tracking, your queries
Citation-source intelligence
AI-crawler traffic analytics
API and MCP access, documented
Every rate with sample size and 95% CI

Agency

$1,199/mo

Up to 25 client domains.

Everything in Monitor, on every domain
White-label reports, your logo
25 client domains included
250 active prompts, pooled across clients (10 each at 25 clients)
Per-client share of voice and sources
Team seats and roles
Priority support

All scans use live data. Scores reflect site-readiness for AI engines, not a guarantee of citations. Read the methodology. See the live dashboard — Aura's own Monitor account, read-only, real data.

Cadence: the 18-check re-scan runs every Monday at 06:00 UTC. Share of voice, prompt tracking and source intelligence sample ChatGPT, Claude, Perplexity and Gemini every day at 08:00 UTC. Your first AI-answer sample starts the moment payment clears and usually lands within a few minutes; if that run fails or hits its time budget, the daily run at 08:00 UTC finishes it, so within 24 hours either way. Until it lands the dashboard says “no sample yet” — it never shows a fabricated number.

What each plan includes

Every limit below is the one the product enforces — the same numbers the dashboard applies.

Feature	Scan $0	Monitor $199/mo	Agency $1,199/mo
Domains monitored	Unlimited one-off scans	1	Up to 25
Active prompts tracked	—	Up to 50 for the domain	Up to 250, pooled across the workspace (10 per client at 25 clients; you allocate)
Named competitors per domain	—	Up to 5	Up to 5
AI engines sampled	—	ChatGPT, Claude, Perplexity, Gemini	ChatGPT, Claude, Perplexity, Gemini
Site re-scan (18 checks)	On demand	Every Monday, 06:00 UTC	Every Monday, 06:00 UTC
AI-answer sampling	—	Every day, 08:00 UTC	Every day, 08:00 UTC
Drift alerts by email	—	Included	Included
Scan history kept	Result link	104 weekly scans (~2 years)	104 weekly scans (~2 years)
Pages audited per domain	Homepage + discovery files	Up to 25	Up to 25
AI-crawler traffic analytics	—	From access logs you upload	From access logs you upload
Team seats and roles	—	—	Up to 25 members
White-label reports	—	—	Included
API keys	—	Included	Included
Exports	Shareable result link	Shareable report links	Shareable, white-labelled report links
Support	Email	Email	Priority email

Over a limit? The dashboard says so at the point you hit it: prompts beyond the limit are not sampled and competitors beyond the limit are not saved. Nothing is charged beyond the plan price. Crawler analytics read the access logs you upload; AI crawlers do not run JavaScript, so no pixel can see them.

COVERAGE

What is measured, engine by engine

The dashboard's engine keys, the product each stands for, the surface actually called, and what is not sampled.

Dashboard key	Product	Measured	What is called	Model route	Web search	Not sampled
`chatgpt`	ChatGPT (OpenAI)	Yes	OpenAI model via API, searching with OpenAI's own web search	`openai/gpt-5.4-mini`	OpenAI web search	The chatgpt.com consumer app, ChatGPT Search as a person sees it, memory or account personalisation.
`claude`	Claude (Anthropic)	Yes	Claude model via API, searching with Anthropic's own web-search tool	`anthropic/claude-haiku-4.5`	Anthropic web search	The claude.ai consumer app, Projects, account context, and Claude's larger models (Sonnet, Opus): answers come from Haiku, Anthropic's fastest model.
`perplexity`	Perplexity (Perplexity)	Yes	Sonar API, which searches the web on every request	`perplexity/sonar`	Perplexity Sonar search	The perplexity.ai consumer app, Pro or Deep Research modes.
`google-ai`	Gemini (Google)	Yes	Gemini model via API, grounded with Google Search	`google/gemini-3.8-flash`	Google Search grounding	Google AI Overviews, AI Mode, and the gemini.google.com consumer app. These are different products and are not measured.
`grok`	Grok (xAI)	Yes	xAI Responses API with the server-side web_search tool	`grok-4.6`	xAI web search	The grok.com consumer app, the Grok panel inside X, and Grok's search over X posts. These are different surfaces and are not measured.
`copilot`	Microsoft Copilot (Microsoft)	No — no supported data source	Not called — no API returns what Copilot shows a person Microsoft publishes no API that returns the citations Copilot shows a person, so any Copilot rate Aura printed would measure a different product than the one the buyer means.	—	No	Every Copilot surface: copilot.microsoft.com, Copilot in Bing, Windows, Edge and Microsoft 365.
`ai-overviews`	Google AI Overviews (Google)	No — no supported data source	Not called — AI Overviews are exposed through no Google API AI Overviews render inside the search results page and are exposed through no Google API, so there is no deterministic, first-party way to sample them. Aura's `google-ai` row measures the Gemini API, which is a different surface.	—	No	AI Overviews and AI Mode in Google Search. The `google-ai` row above measures the Gemini API, which is a different surface.

Where from: Requests are made from Aura's servers in Vercel's default region (US East); no location, language or account personalisation is applied. A project's country setting is used for prompt wording only.
Language: Prompts are asked in the language they are written in; Aura does not translate.
How often: On every paid plan, every active prompt is asked once per engine per day at 08:00 UTC. One answer per prompt per engine per day; there is no re-asking to smooth results.
Citation benchmark: The citation benchmark asks a fixed set of questions derived from your site on every engine once a week and records whether your domain is cited; it is read through the API and MCP. The dashboard's citation rates come from the daily sample at 08:00 UTC, which records every site each answer cites.
Answer length: Each engine may write up to 4,096 tokens, enough for a complete answer; an answer that still runs into the cap is marked truncated on its trial.
Statistics: Every rate is k mentions out of n answers with a Wilson 95% interval. Fewer than 30 answers, or an interval wider than ±15 points, is shown as a count, not a rate.

Changes to what is measured
Date	Engine	Before	After	What it means
2026-10-01	Gemini	google/gemini-2.5-flash, answering from Exa web results	google/gemini-3.8-flash, grounded with Google Search	Before this date the Gemini row was not grounded with Google Search, although this table said it was: OpenRouter served Gemini 2.5 Flash with Exa's search results. Citations recorded before this date reflect Exa's index, not Google's.
2026-10-01	Claude	anthropic/claude-sonnet-5 with Anthropic web search	anthropic/claude-haiku-4.5 with Anthropic web search	Same search backend; a smaller Claude model writes the answer, in line with the cost-efficient models the other rows use. Mention rates either side of this date are not directly comparable.
2026-10-01	All engines	Answers capped at 1,024 tokens	Answers capped at 4,096 tokens; a capped answer is marked truncated	The old cap cut long answers off mid-list, so a brand named late in an answer could be missed. Mention rates after this date can be higher for that reason alone.

Pricing questions

Is the AI visibility scan really free?: Yes. The 18-check scan is free for any domain, with no account and no credit card. You get the full report, every fix, and a permanent shareable link. The email box on the home page sends a courtesy copy of the report; it is not a paywall.
What does Monitor at $199 per month include?: One domain. Every day, up to 50 of your prompts are asked on ChatGPT, Claude, Perplexity and Gemini (via their APIs with web retrieval), and each answer is checked for mentions of you and up to five named competitors, position, sentiment and citations, with a readable excerpt kept as evidence. Every rate carries its sample size and a Wilson 95% confidence interval. The 18-check re-scan runs every Monday with drift detection and an alert email on a major drop. AI-crawler traffic analytics come from access logs you upload.
What does Agency at $1,199 per month include?: Everything in Monitor across up to 25 client domains, with white-label reports, team seats and API keys. The prompt allowance is 250 active prompts pooled across the workspace: with all 25 clients enrolled that is an average of 10 prompts per client, against Monitor's 50 for one domain. You decide how to allocate them.
Can I cancel?: Yes, any time, from the Stripe billing portal (linked from Settings and your receipt) or by emailing support@aaivisibility.com. When Stripe ends the subscription, re-scans, AI-answer sampling and Monitor email stop and nothing further is charged; your saved reports and scan history stay readable. Prices exclude any tax Stripe applies at checkout.
Why does the free scan not show what percentage of AI answers mention my brand?: Because a single run is not a rate. AI answers vary run to run, so a percentage from one measurement session has no statistical meaning. Monitor samples repeatedly over time and reports every rate with its sample size and a 95% confidence interval; no headline number ships without its interval.
Which currencies and payment methods do you accept?: Prices are in US dollars. Payment is by card through Stripe Checkout; card details are entered on Stripe and never touch our servers. Promotion codes can be applied at checkout.

Start with the free scan.

About ten seconds, no signup, every fix spelled out. Upgrade when you want the measurement and the evidence.

Run a free scan