Lapilli

Write firmware once. Run it on your laptop, inspect it in the dashboard. Flash it to the chip.

Get started Repository

The same C source file runs as a native host process and as MCU firmware. Only the port layer changes — selected at build time, invisible to application code.

firmware.c

examples/blinky/main.c

#include "hal/hal.h"

int main(void) {
  lpl_gpio_init();
  lpl_gpio_set_mode(LPL_PIN_A0, LPL_GPIO_OUTPUT);

  while (1) {
      lpl_gpio_write(LPL_PIN_A0, LPL_GPIO_HIGH);
      lpl_delay_ms(500);
      lpl_gpio_write(LPL_PIN_A0, LPL_GPIO_LOW);
      lpl_delay_ms(500);
  }
}

build & run

terminal

# Run on the host — no hardware needed
make TARGET=host examples
./build/host/blinky

# Cross-compile for STM32L4 and flash
make TARGET=stm32l4 examples
st-flash write build/stm32l4/blinky.bin 0x8000000

$ python3 dashboard/run.py —firmware build/host/multi_node — two MCUs, simulation bus, live event log

One API, many targets

Write against hal.h. The build system picks the port directory — no #ifdef soup in shared code, ever.

Test without hardware

All logic runs as plain host processes. The POSIX port is the primary CI target. Hardware is never a prerequisite.

Live dashboard

Visualize GPIO, UART, ADC, and the simulation bus in real time from a web dashboard.

No overhead

C11. No dynamic allocation. No vendor SDKs in shared code. No feature added “just in case.”