Voice-controlled Home Automation from Scratch Using IBM Watson, Docker, IFTTT, and Serverless

on YouTube ↗37m

" summary

A working voice-controlled demo wires a Raspberry Pi mic to a chain of serverless functions that record audio to file and POST it to IBM Watson speech-to-text, feed the text to a trained Natural Language Classifier, look up device IDs in a NoSQL store and publish commands via OpenWhisk serverless actions to an MQTT IoT broker (used with a Wemo-switched light). SMS via Twilio and IFTTT maker channel are supported. Functionality works for simple on/off intents but latency and scale are unbenchmarked and not optimized for high-throughput real‑time use.

AI-written summary · about

" chapters (15)

1. 0:00Introduction
2. 2:00Architecture Overview
3. 15:10Demo
4. 16:02Architecture
5. 17:32IBM Watson
6. 19:03Speechtotext
7. 19:50Natural Language Classifier
8. 20:36Training Data
9. 21:33Classes
10. 22:19Communication
11. 24:24Web Triggers
12. 28:19Extending the Architecture
13. 29:15Testing the Architecture
14. 30:27Technical Recipes
15. 30:52Questions

" related talks

• 34mVoice-Controlled Lights on a Raspberry PiMisael Pérez Chamorro · ElixirConf EU 2026
• 45mVoice Assistants & IoT AutomationDaniel Holmlund · Global IoT DevFest 2017
• 39mVoice Controlled Home Automation in RubyJeff Sacks · RubyConf 2017
• 10mA Day in the Life: An Immersive Data ExperienceDavid Boloker · Velocity Conference Amsterdam 2015
• 12mA Day in the Life: An Immersive Data ExperienceDavid Boloker · Velocity Conference New York 2015
• 01h 25mUsing Nexmo Voice WebSockets with IBM WatsonAaron Bassett, Dave Nugent · Vonage Campus 2019
• 40mIBM Watson ConversationFrancesca Gigante · Codemotion Milan 2016
• 14mKeeping your head cool with serverless voice controlled fanPaco van Beckhoven · J-Fall 2018