AIY Voice Kit Clock 🕰️

A talking clock for people who prefer listening over reading.

Designed for simplicity and accessibility, it speaks the current time or a compact weather update in natural German at the press of a single button. This proof of concept runs on a Raspberry Pi, uses AWS for serverless backend logic and WeatherAPI.com for forecast data, and plays audio via mpg123/ALSA on the Pi.

Purpose

The project was created for a person with limited vision. The goal: a device that works with one button, gives clear spoken feedback, and avoids the complexity of menus or screens.

Features

• One-button interface with multi-press patterns (no screen required).
• High-quality German Text-to-speech via AWS Polly.
• Weather briefings (date, weekday, time, compact forecast).
• Boot-time auto start (systemd service).
• Lean Raspberry Pi setup (no extra Python packages on the Pi; uses mpg123 for playback).

All speech output is German. To maximize voice quality, a generative AWS Polly voice is used.

Audio examples:

• Current time
• Weather forecast

Prerequisites

• Google AIY Voice Kit v1
  Built and tested with the first version of the voice kit. Newer versions will likely work with small adjustments.
• Raspberry Pi 3 + MicroSD card
• Wi-Fi (or LAN)
• AWS account
• WeatherAPI.com API key (free tier works)
• Dev machine: Python 3.12, AWS CLI, Task

Note

Pi-side code is compatible with Python 3.7.3 (depending on the AIY image). Server-side Lambdas target Python 3.12.

Getting Started

Setup Guide

Usage

Operate everything with the single button. Multi-press events trigger actions:

• 1× — 🕰️ Current time
• 2× — 🌤️ Weather forecast + date, weekday & time (may take up to 15s to respond)
• 5× — ℹ️ Instructions
• 6× — 🔧 Self-diagnosis
• 7× — ⏻ Shutdown

(3×/4× are reserved and not used at the moment.)

License

The project is licensed under the Apache License 2.0. See LICENSE file for details.