Most analytics tools price the bottom tier under cost and hope you grow into the margin. Citino does not have to, because it does not measure accounts — it measures categories. One collection of Warehouse Management Software · US runs every day and answers for every subscriber in that category at the same time. The fifty-fourth prompt costs exactly as much whether one brand is watching it or forty are.
So the marginal cost of adding you to a category that is already collecting is $0.00. Starter is $79 because of arithmetic, not because it is a loss leader with a trap further up the page. It is also why your first chart has a line on it on day one. The only thing that grows with you is custom prompts — the questions only you want asked, which are a real engine bill every day, forever. That is what the tiers meter.
Every tier reads the same daily category index. What changes is how many categories you read, how many questions of your own you add to it, and whether the product is allowed to tell you what to do about the answer.
One brand, one market, the whole measurement.
Everything the measurement is. None of the action layer.
For the person who has to do something about it on Thursday.
The tier the product was designed around.
For a team running this across a portfolio.
The whole index, programmatically.
For someone reporting to clients rather than to a board.
Priced per roster, not per client. Three of ten seats used is still $499, and the tenth costs nothing extra.
USD, monthly, billed monthly. Annual and the India price book are two sections down.
On "5 engine slots." The roster today is four: ChatGPT, Gemini and Perplexity, with Claude as an add-on. Starter samples the three. Growth buys five slots against a roster of four, and the fifth is paid for and deliberately empty — when another engine ships an official answer API it lands in Growth at the same price instead of becoming the reason to invent a new tier. Scale takes every engine, now and later. Copilot and Google AI Overviews are on nobody's plan: neither exposes an official answer API. Both still appear in attribution as labelled referral sources.
On Starter's cadence. Every other day is half the answers, so your custom prompts on Starter carry a visibly wider band than your category numbers do. The product prints the wider band rather than smoothing the line and hoping you do not check. If that band is too wide to act on, that is the honest reason to be on Growth — not a feature you are locked out of.
A project is one brand, its competitive set and its GA4 property. A country is a locale the prompt set runs in. A client seat is one client: their brand, their category, their GA4.
Whatever you pay each month, three things hold on every plan: the annual discount, the local price book, and the seat count.
Annual. Pay for the year and take a quarter off. Starter $711, Growth $2,241, Scale $5,391, Agency $4,491. No minimum term on monthly, and switching back is a setting rather than a conversation with someone whose job is to keep you.
The India price book. India is priced in rupees on its own price book, not converted from dollars at whatever the rate was that morning. Billed in rupees, GST invoiced, rupee prices shown on the India page and at checkout.
Seats. Unlimited users on every plan, Starter included. A seat costs us nothing, because the collection is already running — and charging for one reliably ends with a single licence and a screenshot in Slack.
Usage pricing is usually a way of making the sticker price look smaller than the bill. Here is the entire meter, on one screen, so you can check it before you buy rather than after.
Custom prompts, counted as prompts and not as answers. One custom prompt asked across three engines in one country is one prompt — not three, and not ninety a month.
Categories you subscribe to.
Countries, each a locale your prompt set runs in.
Projects, each one brand, its competitive set and its GA4 property.
Client seats, on Agency only.
Category prompts. All 54 of them in Warehouse Management Software · US, every day, on Starter exactly as on Scale. They are a fixed cost shared by everyone in the category, so metering them would bill you for something already paid for.
Answers. 1,785 of them in thirty days in that one category. There is no per-answer, per-engine-call, per-token or per-report charge anywhere in the product.
Seats, logins, dashboards, saved views, scheduled reports.
Reads of your own archive, however far back you go.
Alerts, however many clear their band.
You cannot pause a category prompt. That is a rule rather than a feature nobody got to yet. The collection is one collection: if any subscriber could switch a prompt off, the gap would appear in every other subscriber's chart in that category, and in the public benchmark page anyone can open and check. You can mute a prompt from your own view and never see it again. The sampling carries on without you, which is the only reason the history is worth anything.
Six categories are published right now, with real measured series, no signup and no email field. They exist because the collection runs whether or not anyone in them is a customer, so publishing them costs nothing and proves the index is real before you have given us anything.
Growth, free for 90 days, at full limits, no card. One condition: you tell us where the measurement is wrong.
That is not a loyalty test. It is the thing we cannot buy — a reading of the numbers by someone who knows their own category better than we do — and it is worth more than $249 a month to get it early.
There is no free tier behind it, and the reason is arithmetic. A category is free at the margin; a custom prompt is not, because it is an engine bill every day, forever, on an account that may never convert. Ten bounded places are the honest version of the same offer.
Custom prompts, categories, countries, projects and — on Agency — client seats. Nothing else. Category prompts never count on any plan: they are collected once and served to everyone in the category. Answers never count either — 1,785 of them in a month in one category, at no per-answer charge. Neither do seats, dashboards, exports, alerts or reads of your own archive. You also cannot pause a category prompt, and that one is deliberate: the series is shared, so a pause would put a hole in every other subscriber’s chart and in the public benchmark page. Mute it from your own view instead; the sampling continues.
Because a seat costs us nothing and per-seat pricing quietly makes the data less used. The people who need this number are the writer who owns the page you are trying to fix, the analyst checking whether last month’s move cleared its band, and the founder who wants to know whether it worked. Per-seat pricing reliably ends with one licence, one screenshot in Slack, and a decision made from a picture of a chart that is three weeks old.
Nothing breaks and nothing is silently dropped. Categories, countries and projects are hard limits: adding an eleventh category on Scale asks you first. Custom prompts above your count are queued rather than deleted — they show as over limit, not sampled with the date sampling stopped, and they start again the day you make room. There is no overage line on the invoice, and we do not disable a dashboard to make a point about upgrading.
No. Category history belongs to the category, not to your plan: drop from Growth to Starter and the one category you keep still carries its full collected series. Custom prompts above the new limit stop being sampled — their archived answers stay readable and their old chart stays exactly as it was, with a hard break at the date sampling stopped. Upgrade again and a new segment starts after the break. The two segments are never joined into one trend line.
A buying market and the fixed prompt set that defines it. Warehouse Management Software is one category; HR & Payroll Software is another. Country is a separate axis over the same prompt set. If your market is not indexed yet you can request it, and a new category opens in 5–10 business days: the prompt set has to be written, reviewed, and then collected long enough to be worth showing you.
A free tier is unbounded — every account’s custom prompts are a daily engine bill with no end date — so it ends up either crippled on purpose or an open tap. Ten design-partner places are bounded and buy something specific: Growth free for 90 days at full limits, in exchange for telling us where the measurement is wrong. Meanwhile the benchmark pages are genuinely free, with real measured data and no signup.
Monthly, self-serve, no demo gate and nobody to talk to unless you want to. The category picker opens on the next screen, and if your market is already indexed you will be looking at a populated chart in about a minute.
Most categories already have months of history collected, so your first chart is populated on day one. Ten design-partner places are open — Growth free for 90 days.
Every figure on this page is illustrative and every brand shown is fictional. The prices are real.