OpenHome - AI Smart Speaker Platform
Built WhatsApp Voice Sender and Alarm/Timer abilities
Visit OpenHomeSmart Speaker SDK
Platform
Voice + Alarm abilities
Features Built
Active Discord dev community
Community
Django + Vue.js + Multi-AI
Architecture
The Problem
OpenHome is an open-source AI smart speaker platform that lets developers build custom voice agents. They needed new abilities added to the platform: a WhatsApp Voice Sender module and an Alarm/Timer system, both integrated with their existing voice SDK and shipped within the developer community's release cycle.
Our Approach
Built the WhatsApp Voice Sender as a standalone ability module that hooks into OpenHome's voice pipeline. The module captures voice input via Deepgram, processes it through the LLM layer, and sends voice note responses through the WhatsApp API. Also built the Alarm/Timer ability with natural language time parsing. Collaborated directly with the CTO (Jesse) through their Discord developer community.
Pipeline Breakdown
01 · Collect
- Voice input via smart speaker
- Natural language commands
- WhatsApp message triggers
02 · Process
- Multi-LLM routing (OpenAI, AssemblyAI, OpenRouter)
- Intent parsing for alarms/timers
- Voice synthesis (ElevenLabs)
03 · Act
- WhatsApp voice note delivery
- Alarm/Timer execution
- Smart home device control
Have a similar problem? Let's talk.
← Back to all work