Skip to main content

Manifesto

What we stand for.

These are the principles we use to decide what to build, what not to build, and how. They're not aspirational. They're the floor.

  1. Tenet 1: Tools should be useful for people.

    The simplest test we know: does a person leave better off than they came? Not “did they get something for free.” Not “did they save time we didn't ask them to save.” Useful for them, in a way they would describe themselves. If we can't answer yes, we don't ship it.

  2. Tenet 2: People over output.

    A tool that helps one person do good work matters more than a tool that floods the world with work no one wanted. We're not in the volume business. We're in the help-someone-do-something-they-care-about business.

  3. Tenet 3: Tools, not solutions.

    A solution promises to take the work away. A tool gives you better hands. The first removes you from the loop; the second keeps you in it. We build the second kind because the first kind always ends with someone realizing the work still has to be done — and now they can't do it.

  4. Tenet 4: Capability, not dependency.

    A good tool leaves you more capable than it found you. A bad tool leaves you needing it. We're trying to build the first kind. If you stop using our products and find yourself less able than before, we failed.

  5. Tenet 5: Use AI when it helps; don't pretend it's the work.

    We use it inside our products. We're not coy about it. But the model is the saw, not the carpenter. The thing we sell is what a person can do with our tools — not what the tools can do without one.

  6. Tenet 6: Language is one wall. There are others.

    We started with language because it's the wall we hit hardest, and the one we know best. But the company isn't *about* translation. It's about removing walls between people and the things they could otherwise do. The next product might not look anything like the first one. That's fine. The principle stays.

  7. Tenet 7: Community is the test.

    A tool is only good if the people it's for say so. We build alongside the people who first asked for what we made — they tell us when something works, when it doesn't, and what to build next. We don't ship to abstractions. We ship to people we know.