Priced on answers hosted, not on how hard you work in one month

A token is one unit of AI work, and everything that runs a model costs a fixed, published number of them: 2 for a theme scan, 4 for a question or trend scan, 7 to plan and write an answer, 1 for a question the Agent answers. The price is quoted before the work starts and charged as quoted — never re-priced afterwards because a page turned out to be long. A typical month on Basic is a few theme scans, a dozen question scans, half a dozen trend reads and around thirty-five answers, which lands inside 500 with room to re-run things.

Work stops rather than silently costing you money. You can top up — 100 for €9, 500 for €39, 2,000 for €139 — and bought tokens carry over month to month, while the monthly allowance does not. Nothing publishes, deletes or degrades because a balance ran out; the hub keeps serving everything already on it.

Because an answer keeps working after the month you wrote it. A monthly quota charges you for the burst of effort and nothing for the asset it produced, which is backwards. Twenty-five answers is enough to find out whether this works for your market; five hundred is more than most companies will ever have questions for.

Because charging for the research would mean selling worse answers to smaller companies. Everything that decides whether an answer is any good — your confirmed knowledge, your themes, the real questions, the pre-publish checks — is in Free. What you pay for is scale, branding, conversion, the team around it, and the tokens to run more of it.

Different cost shape. The rest of AnswerLink costs something when you publish; the Agent costs something on every visitor who talks to it. Folding it into a plan would either make the plan expensive for companies who never switch it on, or make us quietly ration it. Forty-nine a month on any plan, included in Managed — and each reply it gives spends one token from your allowance, so a busy hub is visible in the meter rather than in a surprise invoice.

We sit down and build the Company Brain with you, map the Theme Canvas, and get a first batch of answers written, checked and live on your domain. After that it is a monthly session: what came in through search and the Agent, what to answer next, what to improve, and a written report. Scoped per company, because the first sprint is real work on both sides.

Because a second person is a different way of working, not a bigger version of the same one. One person confirming records and approving answers needs no assignment, no tasks and no queue; two people need all three, and those are what Pro adds. Free and Basic are a single seat and priced as one.

No, and we do not sell one. Your hub measures itself: it is on your own domain, so the analytics are first-party, there is no third-party tag for a blocker to strip and no shared cookie to ask permission for. Statistics per answer, engine citations we actually observed, and the questions visitors asked all come from the hub itself.

They are on your domain and they stay there — you export the pages and the hub keeps serving until you point the CNAME elsewhere. Your leads export too. There is no version of this where your content is hostage to a subscription.

Every feature on this site carries its state, and so does this table. Anything marked Planned is not shipped and is not invoiced as though it were.