Audio Server

A minimal FastAPI server responsible for playing WAV audio via aplay. Runs on a host with a connected speaker (Raspberry Pi / any Linux with ALSA). Called by the Homemade Alexa Server after a TTS response is generated.

Requirements

• Python 3
• aplay (from the alsa-utils package)
• Configured ALSA audio output

Installation and usage

pip install -r requirements.txt
./run.sh

The server starts on 0.0.0.0:1234.

Deployment

The server address is configured on the Alexa Server side in config.toml:

[audio_server]
url = "http://YOUR_AUDIO_SERVER_IP:1234"

Endpoints

Method	Path	Description
POST	/api/play	Body: raw WAV bytes; starts playback via aplay; returns immediately with {"status": "playing"}
POST	/api/abort	Stops current playback (SIGKILL on aplay); returns {"status": "aborted"} or {"status": "nothing_playing"}

Behavior

• Playback runs in a separate thread — the server responds immediately without blocking subsequent requests.
• Calling /api/play while playback is already in progress kills the current aplay process before starting a new one.
• /api/abort is called by the Alexa Server when the ESP32 device detects a wake word during response playback.