Emberkit is a software studio being built in the open — a home for intelligent products across many domains, not one. Different industries, one standard of craft.
No spam. One note when the first product is live.
Our vision
Most software companies pick a lane and spend a decade in it. We think that made sense when building anything took an army. It does not anymore. The cost of turning a good idea into a working product has collapsed, and what is left is the part that was always hard: understanding a domain well enough to know what is actually worth building.
So Emberkit is not a product. It is a workshop — a place where many products get made, each one aimed at a different corner of the working world. A tool for a clinic. A tool for a trading desk. A tool for a warehouse, a classroom, a storefront, a studio. On the surface they have nothing in common. Underneath, they share the same spine: careful engineering, an interface that respects your time, and a model doing the work no human should have to do by hand.
We are betting that the next great software company is not vertical or horizontal — it is plural. That one team, armed with the right leverage and an unreasonable standard, can serve a dozen industries properly instead of one industry lavishly. Every product we release makes the next one faster to build. The kit compounds. That is the whole idea in the name.
Right now this page is a promise rather than a product. That is deliberate — we would rather show you something real than something rendered. The first release is being built. When it lands, it will be small, sharp, and genuinely useful to someone on their worst Tuesday. Then we do it again, somewhere else entirely.
An ember is not a fire yet. It only needs air.
Every product we ship starts with a model in the loop — reading the messy input, doing the tedious part, and handing back something a human can act on in seconds. We do not bolt AI on at the end. We assume it from the first line.
The domains look nothing alike. The problems rhyme. Someone is retyping data that already exists, waiting on a report that could write itself, or making a judgement call with half the context. We build the same quality of tool for each of them.
No five-year roadmaps. Each product is scoped so it can be useful to one real person in one real workflow within weeks — then sharpened by the people who actually use it. Slow is a decision, and we have decided against it.
Where we are pointing
Not a roadmap — a direction. Each of these has a workflow somebody still does by hand at 11pm.
Records, reporting, workflow
Signals, risk, automation
Routing, capacity, visibility
Practice, feedback, mastery
Catalogue, pricing, retention
Speed, ergonomics, trust
Search, valuation, paperwork
Drafting, editing, publishing
…and the ones we have not named yet
Leave an address and we will tell you the moment there is something to use — not before.