Buyer terms
What you may ask, what you get back, and what you may not do with it. Last updated 9 October 2026.
What you are buying
Single-choice questions, one to five in a study, shown to developers in the idle window of their AI coding tool and answered with one key. You buy answers, not attention time and not impressions:
You are charged for valid answers. An answer that fails an attention check is not one, is not billed, and is not in your results.
What you may not ask
Every study is read by a person before it goes live. These are the grounds for refusing one, and they are the same list our reviewer works from:
- Harvesting personal data. A question whose purpose is to collect a fact about the respondent rather than an opinion — a name, an employer, an email address, a location finer than a country, or anything that would identify them.
- Political or adult content.
- Deceptive framing: a question that misstates what it is for, or that reads as neutral while being written to produce one answer.
- Questions phrased as feedback about Claude Code or Anthropic. Tickover runs inside someone else's tool and does not sell opinions about it.
- A hidden sponsor. The sponsor name you give is shown with the question, and it has to be the real one.
A rejected study costs you nothing: credit already held for it returns to your balance, and a study still awaiting payment is never invoiced. A first study rejected at cost does not use up your at-cost entitlement.
What you may not do with the results
Results come as one row per answer, with a respondent code, a country, a primary language, an activity tier and an operating system. The code is salted per study, so the same person is a different code in each of them.
You may not attempt to re-identify a respondent, and you may not combine two exports to follow one respondent across studies. You may not sell or pass on the row-level export; the findings drawn from it are yours to publish.
You may not contact respondents. There is no channel to do so, and asking for one in a question is harvesting personal data.
Payment
A study your credit balance does not cover is paid by PayPal before review starts; the amount due is shown on the study page. Credit is an internal balance made of refunds for invalid or unfilled responses and any overpayment. It pays for later studies and cannot be transferred to another account.
A study you submit places a hold on your credits, and the hold is released when the study settles. If a study cannot be filled, the unspent part comes back to your balance.
Founding buyers
For the first five buyers, by arrangement with us by email, a founding study is written with us first and paid when we tell you the panel can fill it, by PayPal or by invoice. Review starts once it is paid, as for every study. If a founding study does not fill within 14 days of going live, we close it and refund the unused part to the account it was paid from. Nothing in the product closes it at day 14; a person does.
Refunds
A rejected study is never charged. Unused credit is refunded on request to the account it was paid from. Write to hello@tickover.dev from the address you sign in with.
What we do not promise
We do not promise a fill time, and we do not promise a full audience: a narrowly targeted study may settle short, and you pay for the responses you got. Answers are opinions given by people in a few seconds, and are not advice you should act on unexamined.
The boring part
Tickover is a trading name of an independent sole proprietor based in Vietnam. Vietnamese law governs these terms. Liability is limited to what you paid for the study in question. Write to hello@tickover.dev and a person will read it.