We’re adaptable full-stack developers — desktop apps, cloud backends, and mobile apps, through to firmware, drivers, and embedded Linux. But one area we know better than most: camera and video systems.
From Emmy Award-winning animal-borne cameras on BBC Blue Planet 2, to custom IP cameras running YOLO on edge AI accelerators, to NVRs handling many channels — we’ve shipped it all.
Full-stack camera product development
Camera products are rarely just about one piece of software. A working product usually needs firmware, operating system work, device drivers, video pipelines, streaming protocols, mobile applications and viewing clients, cloud services, and support tools all fitting together.
That is where our experience is strongest: building across the whole stack, from board bring-up and bootloaders through to mobile apps, cloud storage, and desktop analysis tools.
IP cameras, CCTV and embedded vision
We have developed multiple from-scratch IP camera and CCTV products using chipsets from vendors including Novatek, HiSilicon, iCatch, TI and ChipWrights.
This work has included:
- Embedded Linux, from-scratch RTOS and off-the-shelf RTOS designs
- OS, BSP, device driver and bootloader development
- From-scratch ONVIF servers and customised RTSP server stacks
- Sensor-to-codec video pipelines and IP streaming protocols
- Reverse-engineering ODM camera hardware and replacing it with custom firmware
- Image sensor driver development, including WDR/HDR sensors
- ISP tuning and camera calibration
- Edge AI / YOLO object detection optimised for embedded SoC accelerators
One recent example is a complete IP camera solution based on the Novatek NT98530 SoC. That included a new RTSP and ONVIF server stack running on the Novatek SDK and Embedded Linux, integration with the Novatek codec pipeline, associated device drivers and bootloaders, and YOLOv8 object detection running on the SoC’s built-in CNN accelerator.
Specialist biologging cameras
We also developed several animal-borne biologging camera products from scratch using video SoCs including TI DM368, multiple HiSilicon variants, and iCatch.
These systems had very different constraints from conventional CCTV products. They were battery powered, had to boot quickly, and needed reliable recording in challenging real-world conditions.
The software stack included:
- Bootloaders, Linux and RTOS work
- Device drivers and image sensor setup
- FFmpeg-based recording
- One-second boot-to-recording behaviour
- Desktop applications for analysing biologging sensor data
- GPS post-processing software
- 3D OpenGL visualisation tools
This work provided the animal-borne cameras used on the BBC Blue Planet 2 series, for which we received an Engineering Emmy Award.
NVRs, decoders and storage-heavy systems
We have also developed multiple NVR products across a variety of chipsets.
That work covered:
- Complete OS, BSP, device driver and bootloader development
- Network and storage device drivers
- IP stacks, SAN and NAS protocols
- Video decoder and display device drivers
- Embedded Linux and RTOS environments
- Analogue and IP video sources
- Local codec pipelines and external plugin hardware codecs
- Transcoding
- Video recording and playback stacks
- Optimised video-wall style multi-screen displays
- Related hardware products including network switches, routers and SAN/NAS devices
NVR work is a good example of why end-to-end video experience matters. The hard parts are often at the boundaries: storage throughput, codec behaviour, display pipelines, network reliability, and how the UI behaves when many live streams are active at once.
Video encoders and streaming pipelines
We have worked on multiple video encoder products using SoCs including i.MX6, Qualcomm Snapdragon 820 and ChipWrights devices.
This included BSP, OS and device driver work, plus optimised video transcoding pipelines for splitting input IP and analogue video streams into multiple output resolutions and formats.
On the streaming side, we have built systems using GStreamer, FFmpeg, MediaMTX and custom pipeline stacks, including:
- Gateway devices bridging local IP cameras to cloud storage
- WebRTC live viewers
- Amazon Kinesis Video Streams integration
- Pipelines connecting local IP camera streams to Android and iOS mobile apps
Where we can help
If you are building a camera product, video pipeline, edge AI system, NVR, encoder, biologging device, or cloud-connected video product, we can help across the stack.
Typical areas include:
- Firmware, bootloaders, BSPs and embedded Linux
- Device drivers and image sensor integration
- Sensor to codec video pipelines
- Customisation of SoC vendor SDKs
- RTSP, ONVIF, WebRTC and cloud video streaming
- GStreamer, FFmpeg and custom media pipelines
- Local and/or cloud video recording & playback
- YOLO / edge AI on embedded accelerators
- Mobile, desktop and cloud applications around video products
- Debugging and improving existing camera/video systems
Camera and video products often fail because one layer was built without understanding the constraints of the others. Our strength is joining all of those layers up: firmware to cloud and customer-facing apps.