We build software that keeps working: intelligence on the device, data under the user’s control, and products that stay useful when the network does not.
Our work
Our products cover filmmaking, photography, accessibility, music, and robotics. Building and maintaining them gives us direct experience with the work our clients need: on-device AI, offline workflows, hardware communication, and interfaces that fit the device.
Captain SuperSight brings on-device vision to an accessibility app. Cinema Slate supports offline production work. BT Cam connects phones and computers to compatible cameras over Bluetooth.
Explore all our products or see how that experience translates into engineering services for your team.
How we build
Offline operation. Essential workflows and local data should remain available when the connection drops. Sharing and synchronization support the experience where needed.
Local processing. We keep computation and data on the device when the task and hardware allow it. AI still needs an honest assessment of model quality, memory, response time, storage, and battery use.
Privacy by design. We ask what information a feature needs before deciding what to collect or send. We minimize collection, make data flows explicit, and give users meaningful control.
Maintainable code. Flutter and Dart let us share logic and interfaces across platforms. Native integrations handle operating-system features and hardware where required. We choose the split around the product and the team that will maintain it.
Open source. We share reusable libraries that also support our products. Universal BLE powers Bluetooth connectivity in BT Cam and Cinema Slate; Nearby supports Cinema Slate’s local synchronization. Developers can inspect the code, evaluate it, and contribute. Each project retains its own license and terms.
Working together
- Define the task and constraints. Agree who uses the product, what it must do, target hardware, data flows, offline requirements, and measurable success criteria.
- Evaluate a working prototype. Test the uncertain parts early: model quality, response time, resource use, or hardware communication. Use the results to choose an approach and scope.
- Integrate and test. Build the feature into the application, test on target devices, and cover the offline, reconnection, and failure paths users will encounter.
Our principles guide how we build: local processing where practical, clear data flows, maintainable code, and reusable open source. Read how we build.
Our expertise
On-device intelligence, privacy-conscious engineering, reliable offline workflows, cross-platform apps, and software that connects to hardware. Explore our specialties through the apps we build and maintain.
Offline-first
Keep essential work available in the field, on a film set, or wherever reception is unreliable. We design local data and workflows first, then add sharing, synchronization, and recovery where the product needs them. Local processing is also a privacy decision: data that stays on the device does not have to be sent, stored, or protected on a server.
Cinema Slate logs takes and keeps shot history on the device, with local sync between nearby devices. Captain SuperSight runs its AI features locally after model setup. Every app we ship follows the same approach. Operating from the EU, we hold ourselves to the standards set by the GDPR and by ethical engineering practice: we collect only what a feature needs, make data flows visible, and give people control over their information. Local storage still needs protection, and where a server is genuinely required we make the transfer explicit.
Edge AI
Bring intelligence into an app without sending each request to a server. We build camera-based AI features and evaluate other uses, such as private document search and local assistants, against your data and target hardware.
Captain SuperSight describes scenes, checks user-defined visual conditions, and speaks the results. Inference runs on the device after model setup. That work gives us direct experience balancing output quality, response time, memory, storage, and battery use.
Local device connectivity
Connect an app to cameras, accessories, and robots. We build around the device protocol, permissions, connection lifecycle, and reconnection behavior, then test on the hardware people will use.
BT Cam controls compatible cameras over Bluetooth. Cinema Slate connects to timecode equipment and nearby devices. UniversalBle Explorer helps developers inspect Bluetooth devices, while Romo and Romo Control support the compatible robot platform. Our reusable libraries carry that experience into other products.
Universal BLEBluetooth Low Energy for Flutter with detailed documentation, built-in permission management, and central and peripheral APIs.
Universal Bluetooth ClassicBluetooth Classic discovery, pairing, HID, and iOS External Accessory integration for Flutter.
NearbyPeer-to-peer discovery and data transfer over Wi-Fi/LAN and Bluetooth Low Energy through one Flutter API.
WiFi DesktopDesktop Wi-Fi connection management, network scanning, and interface control for Dart.
Romo SDKAn iOS SDK for Romo hardware, character animation, and robotic interaction.
Cross-platform
Build the product once and ship it where users are: iOS, Android, desktop, and the web. We share interface and business logic in Flutter and Dart, then use native code for the platform features that need it, including permissions, audio, Bluetooth, and background execution.
BT Cam and Cinema Slate run on iOS, Android, Windows, macOS, and Linux, and as web apps. UniversalBle Explorer covers iOS, Android, and the web; Tunedeck and Gesturedeck run on iOS and Android. The split between shared and native code follows the product and the team that will maintain it, and each platform is tested on real devices.
For reusable building blocks, explore our open-source libraries and commercial SDKs. For the platforms, frameworks, and tools we use, explore our technologies.