How we build, and how we keep building.

Private where it counts

Wherever the work can happen on your own device, that is where we do it. EverHeard captions speech and recognises sounds entirely on your phone or Mac: no account, no server listening, nothing recorded.

Not everything can work that way, and we would rather say so than imply otherwise. Some tasks need more capability than a phone has, and when you ask for one, that request leaves the device. In those cases we hold to three rules: send the least that will do the job, strip what identifies you where we can, and keep nothing beyond what you asked us to keep. The result comes back to your device or to storage we control — not to an advertising market, and not into training data.

We apply the protections the technology allows, and we tell you where the line falls in each product. What we will not do is promise a level of privacy the software cannot actually deliver.

Accessible by design

We design for the people most often left out, and everyone else benefits. Captions that keep the speaker's own language rather than translating it away. Alerts you can feel instead of hear. Interfaces that respect your text size, your contrast and your preference for less motion.

Accessibility is not a feature we add at the end. In our flagship product it is the entire point.

Honest about limits

Software that is confidently wrong is worse than software that admits doubt. Each of our products says this in its own way, because they fail in different ways: EverHeard must never fabricate certainty for someone who cannot check it against what they heard; ResumeCraft_AI turns gaps into questions rather than inventions; Myth MacroMKR promises only what its code enforces, shows the evidence behind each recommendation, and says plainly when the data cannot answer.

The same discipline governs how we describe our own work. Every status on this site is the real one. A product approved but not yet released says exactly that. A feature still being built is never described as available. We publish no user, download or revenue figures, because we would rather say nothing than imply something.

Built to last, and never a trap

We build so our products keep working when the network doesn't, recover from their own failures, and don't depend on somebody else's pricing decisions to stay alive. Where a product can run entirely on your device, it does.

And nothing we make locks you in. What you create with our software stays yours, in a form you can take with you and read without us.

Funded so the work can continue

We are a for-profit company, on purpose. We would rather charge honestly for what costs money than quietly sell the person using our software.

Some of what we make belongs to the community rather than to a price list. Where the good a product does outweighs what we could earn from it, we commit to keeping it free — and we fund that commitment deliberately, through grants, sponsorship, institutions, donations, and the revenue of the products that do charge. Working with funding partners is how a free promise is kept rather than merely made.

The test is the work, not the wallet. This isn't about who can afford what; it's about what shouldn't be charged for in the first place.

So we won't tell you everything we make will be free forever. We will tell you, for each product, what it costs and who pays — and where we have committed to keep something free, we name it and we honour it.

How the work gets done

Product decisions, architecture, domain logic and validation on real devices are ours. Code implementation is AI-generated under direction, then reviewed and corrected. It lets a small studio hold a high standard across several products, and we say so plainly rather than implying a larger team.