Question 1

How do I get started?

Accepted Answer

Enter your domain — your first visibility score needs no signup and runs in about 30 seconds. To track over time, create an account; setup pre-fills your prompts and competitors in under 15 minutes. Start from the homepage.

Question 2

What's the difference between the free audit and an account?

Accepted Answer

The public audit is a one-time, no-signup snapshot of where you stand right now. An account turns that into ongoing tracking: multi-run trends over time, competitor comparison, source-gap analysis, and alerts when something moves.

Question 3

Do I need a credit card to try it?

Accepted Answer

No. The public audit needs no account and no card. You only enter payment details when you choose a paid plan — see pricing.

Question 4

How long does setup take?

Accepted Answer

Under 15 minutes. We auto-discover your competitors and pre-fill a buyer-intent prompt set from your domain, so you mostly review and approve rather than start from a blank page.

Question 5

How many prompts should I track?

Accepted Answer

Enough to cover the real buyer questions in your category — typically 20–50 to start. Each plan includes a per-brand prompt allowance (Starter 30, Growth 100), and you're never billed per run. See pricing.

Question 6

How is this different from an SEO tool?

Accepted Answer

SEO measures blue-link rankings; CitedOS measures whether AI assistants recommend you inside their answers — a different surface, different sources, and a different playbook. Read the GEO guide for the full distinction.

Question 7

How is the Visibility Score calculated?

Accepted Answer

It is the share of tracked AI answers in your category that mention you, sampled across the AI engines you track. Every score ships with a confidence range — click any number to read the raw answer behind it.

Question 8

Why does every number have a range?

Accepted Answer

Because AI citations swing 40–60% month to month — one source's ChatGPT share fell from ~60% to ~10% in two weeks. We run each prompt and engine multiple times and report a mean with a 95% confidence interval, so you act on signal, not noise. See confidence ranges.

Question 9

What is multi-run sampling?

Accepted Answer

AI answers are non-deterministic — ask twice, get two answers. We run each prompt on each engine at least five times and aggregate to a distribution, so a metric reflects how the engine typically answers, not a single lucky or unlucky roll.

Question 10

What are model-version markers?

Accepted Answer

We tag every datapoint with the exact engine model version it came from and draw a marker on your trend lines when that version changes — so a jump caused by a new model release never gets mistaken for a real shift in your visibility. See methodology.

Question 11

What do share of voice, position, and sentiment mean?

Accepted Answer

Share of voice is your mentions divided by all brand mentions in the answer set; average position is where you land when mentioned; sentiment is a 0–100 read of how you're framed. See the glossary for every metric.

Question 12

Isn't a tool finding the same sources just correlation, not causation?

Accepted Answer

Yes, and we treat it that way. We compare cited versus not-cited sources that appeared for the same query, and we separate selectability (on-page signals) from discoverability (off-site authority) — so the playbook reflects real differences, not confounders. See source gap and playbook.

Question 13

Which engines do you measure?

Accepted Answer

ChatGPT, Gemini, Claude, Grok, DeepSeek, Google AI Overviews and Perplexity — via official APIs and licensed SERP data, never by scraping the consumer apps. Starter and Agency Starter track ChatGPT; Growth and Agency Growth include ChatGPT, Gemini, Claude and Google AI Overviews; any engine your plan leaves out is a $50/mo add-on. Engines and data explains how each one is accessed.

Question 14

Is the ChatGPT data official?

Accepted Answer

Yes — it comes from OpenAI's official web-search API, which we label an API / proxy surface because it is an honest approximation of the consumer ChatGPT experience, surfaced clearly rather than overclaimed.

Question 15

Do you scrape ChatGPT, Gemini, or Claude?

Accepted Answer

No. We never scrape the consumer chatbots — that violates their terms and breaks constantly. We use each provider's official API, and the only pages we scrape are the public source pages an engine already cited.

Question 16

Do you support Perplexity or Meta AI?

Accepted Answer

Perplexity yes, Meta AI not yet. Today we measure ChatGPT, Gemini, Claude, Grok, DeepSeek, Google AI Overviews and Perplexity, via official APIs and licensed SERP data; Starter and Agency Starter track ChatGPT, Growth and Agency Growth include ChatGPT, Gemini, Claude and Google AI Overviews, and any engine your plan leaves out is a flat $50/mo. Meta AI and Copilot have never been measured; we add an engine only when we can access it through a compliant official source.

Question 17

How fresh is the data?

Accepted Answer

Paid plans refresh daily, and answers carry a freshness timestamp. We respect each engine's own update cadence (ChatGPT and Bing refresh roughly every 24–72 hours) rather than re-running faster than the underlying data can actually change.

Question 18

How do you find the source gap?

Accepted Answer

We collect the URLs each engine cites in answers about your category — about 90–95% are third-party sites, not your own — then scrape, embed and compare those pages to learn the citability signature and show exactly which sources you're missing from. See source gap and playbook.

Question 19

What is an Agent Experience Platform (AXP)?

Accepted Answer

An AXP is middleware that detects AI agents and crawlers at the CDN edge and serves them a clean, machine-readable version of your pages, while human visitors keep the normal site. It's an emerging category aimed at one problem: making JavaScript-heavy pages easy for engines to parse and cite. See what an AXP is.

Question 20

How does an Agent Experience Platform work?

Accepted Answer

In four steps: it detects AI user-agents (such as ChatGPT-User or PerplexityBot) at the edge, routes them away from your normal origin, serves server-rendered HTML with the facts up front and no JavaScript dependency, and leaves search-indexing crawlers untouched. Humans still get your original site on the same URLs. The full walkthrough is in our AXP explainer.

Question 21

Is serving AI agents different content the same as cloaking?

Accepted Answer

Only if the substance differs. Google prohibits cloaking — showing engines different content to manipulate rankings — but permits making the same content more accessible. An AXP that reformats identical facts into cleaner HTML stays on the right side of that line; one that changes the claims does not. We unpack the test in the AXP explainer.

Question 22

Do I need an AXP to get cited by AI?

Accepted Answer

Usually not first. Server-rendering your key facts and adding structured data solve most of the parsing problem without edge middleware. (You'll also see llms.txt suggested — it's cheap to publish, but no major engine currently reads it, so don't count it as a fix.) An AXP is an enterprise answer for very large or JavaScript-bound sites — reach for it last, not first.

Question 23

Does CitedOS include an Agent Experience Platform?

Accepted Answer

No — CitedOS measures and diagnoses, it doesn't reformat or serve your pages. We show where you're cited across every engine on your plan, the agent traffic hitting your site, and which sources you're missing from; an AXP is a separate, complementary tool that changes what agents receive. Start by measuring where you stand.

Question 24

Is there a free plan?

Accepted Answer

The public audit is free and needs no account — a no-signup snapshot of where you stand. Ongoing tracking is paid; brands start at $99/mo and agencies at $250/mo. See pricing.

Question 25

What do the paid tiers include?

Accepted Answer

Starter ($99/mo) tracks ChatGPT across 30 prompts; Growth ($399/mo) includes ChatGPT, Gemini, Claude and Google AI Overviews across 100. No tier bundles the whole roster — whatever your plan leaves out is a flat $50/mo per engine, on every paid plan. Every paid plan includes unlimited seats, daily refresh and confidence ranges on every metric — no measurement credits to ration — plus the source-gap engine, a prioritized playbook, page audits and MCP access; Growth and Agency Growth add the agent-draft lane. See pricing.

Question 26

Do you charge per engine or per model?

Accepted Answer

Per engine, above what your plan includes — a flat $50/mo each, on every paid plan. Starter and Agency Starter include ChatGPT; Growth and Agency Growth include ChatGPT, Gemini, Claude and Google AI Overviews; you add the rest only if you want them. We never charge per model, and since a prompt is one buyer query we run across your plan's engines multiple times, you're never billed per run. See pricing.

Question 27

Do you have agency or multi-brand pricing?

Accepted Answer

Yes. Agency plans bundle multiple client brands under one account with per-client prompt allowances, per-client dashboards, white-label reports and centralized billing — from $250/mo for 3 client projects, scaling by client volume. See pricing.

Question 28

Can I cancel anytime?

Accepted Answer

Yes. Monthly plans are month-to-month with no long-term contract — pay yearly instead and you get twelve months for the price of 10 on every self-serve plan except Agency Growth, which bills monthly only. Either way you're billed through a merchant-of-record that handles cross-border tax and payments, and you can cancel whenever you like.

Question 29

Is there an API or MCP server?

Accepted Answer

Yes. Every paid plan includes MCP access, so your own Claude or ChatGPT agent can call get_visibility and get_source_gap directly. See the docs.

Question 30

Do you need access to my analytics or CRM?

Accepted Answer

No. CitedOS measures public AI answers and public cited pages, so it never requires access to your private analytics, CRM or customer data to produce a visibility report.

Question 31

Where is my data stored?

Accepted Answer

On managed Postgres (Supabase) in a single, access-scoped system. We store the prompts you track, the AI answers and citations we collect, and your account — nothing from your private systems.

Question 32

Is my workspace data isolated from other customers?

Accepted Answer

Yes. Every query to your data runs through an access-scoped authorization boundary so one workspace can never read another's, and we are adding row-level security on customer tables as an added layer of defense-in-depth.

Question 33

Can I delete my brand or my data?

Accepted Answer

Yes. You can delete a brand and its tracked prompts, answers and metrics from your workspace at any time, and you can close your account entirely on request.

Question 34

Do you sell or share my data?

Accepted Answer

No. Your tracked prompts and results are yours; we don't sell them or share them with other customers. The only outbound requests we make are the official engine API calls and the public source-page fetches needed to build your report.

Questions, answered honestly.

Getting started

Methodology & honesty

Engines & data

Agent experience & AXP

Pricing & plans

Privacy & security

The fastest answer is your own audit.