I built a physical, touch-screen Desktop Voice Agent using t... by Stuart GardollI built a physical, touch-screen Desktop Voice Agent using t... by Stuart Gardoll

I built a physical, touch-screen Desktop Voice Agent using t...

Stuart Gardoll

Completed work

AI Automation

Data Engineer

Cal.com

ElevenLabs

I built a physical, touch-screen Desktop Voice Agent using the Pimoroni Presto and ElevenLabs Conversational AI.

Unlike standard smart speakers, this is a custom-engineered solution designed for high-stakes productivity.

- Hardware-Software Synergy: Integrated MicroPython-based IoT hardware with a Node.js backend.

- Agent Orchestration: Configured a "Non-Nonsense" Executive Assistant persona using ElevenLabs' LLM routing (GLM 4.5 & GPT-4o Mini).

- Tool Use (Function Calling): Implemented real-time calendar management via Cal.com and Google Calendar API integration.

- Low-Latency Performance: Optimized voice response times for natural, back-and-forth human-machine interaction.

- Tech Stack: Python, Node.js, ElevenLabs API, Cal.com, Pimoroni Presto (Hardware). https://youtu.be/xKbjJ2QY9Gc

Like this project

Completed work

Posted Dec 24, 2025

I built a physical, touch-screen Desktop Voice Agent using the Pimoroni Presto and ElevenLabs Conversational AI. Unlike standard smart speakers, this is a cu...

Likes

Views