MANIFESTO
We would rather say nothing than say something untrue.
Every marketing tool built in the last two years can write a paragraph about your business. Almost none can tell you whether it is true.
That is not a small gap. A made-up review is a lie with your name on it. A wrong phone number is a customer who never arrives. An offer with no end date is a promise you did not make. The tool produces it, and you carry it.
So we built the constraint first and the writing second. Every fact about your business sits in a state you set. Anything not verified is held out of drafts — not softened, not hedged, held out. If that leaves a sentence unwritten, the sentence goes unwritten.
A tool that cannot say “I don't know” will eventually say something worse.
This costs us things. It makes onboarding slower, because you confirm what is true before you get output. It makes our marketing quieter, because we will not put a testimonial on our own homepage until a real one exists.
We think that trade is the product. Anyone can generate more. Almost nobody is willing to generate less.
This is why we favor specifics over adjectives. We'd rather show three real customer quotes than a hundred synthetic ones, flag a claim as unverified than dress it up as fact, and cite a source than assert a number. None of this is glamorous. It just happens to be the only version of the product we'd trust with our own name on it.
Questions people ask us about this
Doesn't refusing to guess make the product less useful?
Sometimes, yes — in the moment. But a tool that guesses confidently trains you to stop checking its work, and eventually that costs you more than a slower onboarding ever would.
How do you decide what counts as 'unverified'?
If a claim can't be traced back to something you gave us, a public source we can cite, or a number we computed ourselves, it's marked unverified rather than stated as fact.
Why not just let AI fill in the gaps and let users fact-check?
Because most people won't fact-check a paragraph that reads confidently. Shifting that burden onto users is how misinformation quietly becomes marketing copy.
Does this make onboarding slower?
Yes. We ask more upfront because we'd rather ask than assume. It's a deliberate trade, not an oversight.
Is this just a positioning statement, or does it change the product?
It changes the product directly: outputs are constrained by what's provided or verifiable, and anything else is flagged instead of invented.