echolocation

Assistive hardware that connects to a keyboard, speaks each key aloud, and forwards held keys to a computer as a standard USB or Bluetooth keyboard.

Built for the M5Stack CoreS3. See spec.md for full product requirements.

Get started on your own hardware

1. Assemble the device

You need a CoreS3, battery bottom, USB host module, audio module, and a FAT32 microSD card with speech files. Part list: firmware/README.md.

2. Build and flash firmware

Install PlatformIO, then from the firmware directory:

cd firmware
pio run -e m5stack-cores3 -t upload

Connect the CoreS3 with a USB-C data cable. The first build downloads dependencies and may take a few minutes.

If upload fails, long-press RST until the screen goes blank and try again. See firmware/README.md for details.

3. Load speech onto the microSD card

Generate WAV files with Piper TTS — see scripts/generate-tts/README.md. Copy the audio folder to the root of a FAT32 microSD card, insert it into the CoreS3, and reboot.

Releases

Pre-built firmware is published on GitHub Releases. Each CoreS3 zip includes bootloader.bin, partitions.bin, and firmware.bin plus flashing instructions.

Repository layout

firmware/           CoreS3 firmware (PlatformIO / Arduino)
scripts/generate-tts  TTS audio generation for the microSD card
assets/             Audio manifest for speech generation
spec.md             Product specification (source of truth)
test-plan.md        Manual test checklist