We're a London-based team, building agentic AI systems that disappear into the products people already love.
Every new AI product asks users to learn a new interface. We think that's backwards. The AI should adapt to the interface — not the other way around.
We don't ship chat widgets or AI sidebars. We embed intelligence into the flows that already exist — making them faster, more predictive, and less error-prone.
We use off-the-shelf models, open-source tooling, and cloud-native infrastructure. Low cost today, hyper-scalable tomorrow. No moat in the hardware — the moat is in the integration.
We build lean. Two people means fast decisions, tight feedback loops, and no organisational drag. We ship early, measure ruthlessly, and iterate.
Most AI tooling is built for the desktop. Most people live on their phones. We optimise for mobile-first: low latency, small models, battery-aware inference.
Deep expertise in both surfaces where people spend their time — and the infrastructure that connects them.
We work from London, England. Remote-first when the work calls for it, but rooted in a city with one of the world's strongest AI and startup ecosystems.