Engineering
Since our founding in 2003, we have built software, developed algorithms, researched core technologies, and operated systems. Some systems have been running stably for more than 20 years.
Distribution
Starting with development and service operations for the Japanese market, we have distributed apps worldwide since 2016.
- Localized and maintained in 39 languages
- Built and operated the store distribution pipeline
- Introduced ad mediation
- Measured and analyzed usage
39 languages / payments across 52 countries and regions and 36 currencies (since 2016)
Core Stack
Security 2017–
Mobile Distribution 2016–
Embedded & OS 2008–
Networking & Protocols 2004–
Business & Commerce Systems 2004–
Imaging, Documents & Language 2003–
Technology Decisions
We have worked on mobile development since 2009, the iPhone 3GS generation, and have distributed our own apps since 2016. The starting point was what the iPhone made possible: a high-performance device that fits in one hand, connected to the network, able to deliver a service anywhere in the world.
What we build comes back to us as usage data. Technology and UI/UX can be evaluated continuously in the real environment, and distributing our own apps gives us a scale of validation that individual projects rarely provide. What we learn there feeds back into design work outside mobile as well.
In 2016 we compared Xamarin, Cordova and React Native at the implementation level, and Flutter has been evaluated since its public release. Today our main path is native implementation in Swift and Kotlin, with SwiftUI in production use from the iOS 15 generation.
Writing close to the OS and the hardware lets us tune both measured and perceived speed in rendering and camera work, and use new OS features or C/C++ libraries without waiting for an abstraction layer. Long-running apps also stay clear of breaking changes on the framework side.
Device Control
Barcode-Based Device Readout
Readings from an instrument are collected over three paths: BLE, Wi-Fi and barcode.
The device sends its measurements wirelessly and also shows them as a barcode on its own display. The app reads that with the camera and lists each value with the time it was captured.
All three paths are handled by one app, so the values arrive on the same screen whichever route they take.
On-Device AI
On-Device AR Recognition
Objects in the camera view are recognized on the device and labelled in AR space.
Recognition runs on the device itself. Each label is anchored to the position of the object, so the text stays with it as the camera moves.
Names and their confidence appear as they are found. Nothing leaves the device.