Bring your own keys, or use ours.
AI assistants are where recommendations happen now. CitePrism tracks whether ChatGPT, Claude, Gemini, Perplexity, Google AI Overviews, AI Mode and Copilot recommend you, who they recommend instead, which sources they trust, and what to do about it.
Free to start. Pro from $29 a month.
Listens to
ChatGPT, Gemini and Perplexity can be read from the apps people use, so you see the answer a person sees. Google AI Overviews, AI Mode and Copilot have no API, so that is the only way to read them.
Product
Visibility, share of voice and citations across every assistant, built from the real answers rather than a score pulled out of the air. Every metric is one click from the words it came from.
Open any run and there is the full answer, word for word, with your brand and your competitors highlighted and the sources it leaned on listed underneath. Nothing is summarised away.
How often you are named, how often you are named first, how that differs by question and by engine, and whether it moved since last period. Click any number and you land on the answers behind it.
The sites the assistants cite when they talk about your category: yours, your competitors, review sites, communities. You learn who is shaping the story, and whether you are in it.
Add the brands you lose deals to and see how often they show up in the very same answers as you. Share of voice from one set of questions, with no extra work.
Guidance
01
Every question where a rival is named more often than you, worst first, with the sources those answers cited and the rival pages the engines link to.
02
What the engines searched for before they answered. The words they add to your prompts are the modifiers buyers use. Track any search as a prompt in one click.
03
Every week, a short action list written from your last 30 days of answers: what to create, what to improve, where to get listed, where to show up. Each item is tied to the prompts it helps.
04
Every Monday, per brand: visibility and share of voice with the change since last week, new gaps and searches worth tracking. Opt out per email type, never all or nothing.
How it works
01
Paste your website. CitePrism drafts the questions your buyers are likely asking and the competitors they compare you with. Keep what fits, change the rest. Ask for more prompt ideas any time.
02
CitePrism asks every engine on a steady schedule, from the apps people use where it can, and keeps each answer. You set a daily spend limit, so the bill stays predictable.
03
Within days you can see which assistants recommend you, where a competitor wins, and which pages earn the citations. Every number links to the answers behind it, so you know what to fix and can prove it moved. Gaps and the weekly digest tell you what to fix first.
What it measures
Every number comes from answers CitePrism collected, on a schedule you control, with the sample size shown right next to it. Small samples are marked as small, not dressed up as trends. Every rate greys out below ten answers. Read it as a clear, repeatable signal about how the models talk about your category.
No AI provider exposes what people actually type. Anyone selling "real AI search traffic" is guessing from referrals. CitePrism will not pretend otherwise, and every definition it uses is documented on this page before you trust a number.
Pricing
Every engine and every feature on your own keys for nothing, or let Pro run it on ours.
Free
$0 forever
No card needed
Every engine and every feature, on your own keys.
Pro
7 day free trial$29 a month
$299 a year saves you 2 months
Runs on our keys, so there is no API console to set up.
Card required. Nothing charged for 7 days.
Custom
Contact
Quoted in writing
For teams that need more room than Pro, or a purchase order instead of a card.
Questions
Short answers here. Each one links to the page that goes deeper.
AI visibility is how often, how prominently and how favorably AI assistants mention your brand when people ask them about your category. Where search rankings measured your position on a results page, AI visibility measures whether you are inside the answer at all. It is tracked with a fixed set of buyer questions, sampled on a schedule, and turned into visibility rate, share of voice, rank and citation metrics.
Read the full definitionWrite 15 to 25 questions the way a buyer would ask them, ask each one several times in fresh chats, and record whether you were named, in which position, which competitors appeared and which sources were cited. One answer tells you almost nothing because assistants answer differently every time. CitePrism runs that loop for you every day and stores every answer.
The step-by-step methodNo, and neither can any other tool. Conversations with assistants are private and no provider exposes them. Every AI visibility tool measures by sampling: asking the assistants your category questions repeatedly and analyzing the answers. CitePrism says so plainly, shows the sample size next to every number, and greys out rates below ten answers.
What these tools actually measureCitePrism tracks seven engines: ChatGPT, Claude, Gemini, Perplexity, Google AI Overviews, Google AI Mode and Microsoft Copilot. The four with a public API are sampled through the API, and the others are read from the product itself so you see the answer a person sees. Every engine is available on the free plan with your own keys.
ChatGPT brand monitoringOf all the brand mentions across the sampled answers to your prompt set, share of voice is the fraction that are yours. CitePrism computes it per engine and overall, next to visibility rate, average rank and first-mention rate, with you and your competitors scored against the same questions on the same schedule.
LLM share of voice trackingBecause they are the pages shaping your category. When an assistant answers with citations, it names the review sites, communities, documentation and competitor pages it trusts. CitePrism collects those citations from every sampled answer and groups them by category, which turns a visibility problem into a list of places to show up.
AI citation trackingThey track the same thing. CitePrism differs in what you can see and what you pay: every chart clicks through to the raw answers, every metric definition is published, the cost of every run is on screen, and sampling can run on your own keys at provider price. Profound wins on managed onboarding, SSO and executive reporting; Peec AI and Otterly win on being polished, simple products that someone else maintains. The comparison pages say so in their own rows.
Side-by-side comparisonsYes. The free plan has every engine and every feature on your own API keys, with no card and no expiry. Pro is $29 a month or $299 a year and runs on our keys with daily usage included. Sampling on your own keys costs about five cents per prompt across the four API engines on their default models, paid to the providers.
Pricing and worked examplesAdd your brand and within minutes you know where you are recommended, where a competitor wins, and what to fix first.