inxware is an open source embedded software component library written in C, maintained by inx limited. It gives embedded product teams a set of pre-integrated, tested components and a hardware abstraction layer, so a product can move between microcontrollers, single-board computers and PCs without rewriting its application layer.
| Repository | What it is |
|---|---|
ert-components |
The main component library and event runtime (eRT). Build guide, porting guide and edge-ML guide live in its docs/. |
ert-build-support |
Toolchains and C standard libraries required by the eRT build system. |
ert-contrib-middleware |
Third-party contributed libraries and middleware used in eRT component builds. |
apps |
Example and default applications built with the Lucid no-code IDE. |
- Architectures: Arm (Cortex-M, Arm32, Arm64), x86_64, RISC-V, Xtensa
- Operating systems: Linux (Debian, Ubuntu), Android and the Android NDK, Windows, FreeRTOS, Arduino, bare metal
- Hardware: Raspberry Pi, Rock Pi and Radxa, Espressif ESP32 / S3 / C3, NXP Kinetis, STM32, Android devices, x86 PCs
Full detail in the ert-components supported platforms table.
The ML components are public, including image and SVM inference with working TensorFlow Lite examples. The inxware edge machine learning guide covers model formats and hardware-specific compilation for Nvidia TensorRT, Hailo, DeepX, Qualcomm, Rockchip, NXP eIQ and Arm Ethos-U.
inxware is built and maintained by inx limited, a UK embedded systems company that designs, builds and deploys connected products for OEMs and service providers.
- Open source at inx — what inxware is and how it fits
- Technologies we build with — embedded, connectivity and AI
- Edge AI development — from prototype to production
- inxware.io — the full platform, including the Lucid no-code IDE