"Hey Mycroft": Getting Started with the OSS Home Assistant
summary
Open source Mycroft runs speech recognition and synthesis locally and claims privacy by default, deployable as a Raspberry Pi image or on Linux (Docker/Android/Windows ports in progress). Demos include a hello-world, a Twitter hashtag reader that fetches five tweets, and a dialog-style knock‑knock skill; skills are Python modules with vocab/dialogue files and registered as git submodules for PR into the skills repo. Hardware options: Mark I ~£/€/$150 or a Pi build with an omnidirectional mic/speaker (~$40); limitations are slower updates, Pi latency, and reduced functionality offline.
AI-written summary · about
chapters (10)
- 0:00Introduction to Mycroft
- 1:23Architecture and Privacy
- 3:23Device Setup and Deployment
- 6:23Core Components
- 7:30Developing Custom Skills
- 12:27Code Structure and Repos
- 16:10Implementing Example Skills
- 21:57Conversational Logic
- 29:09Testing and Debugging
- 33:51Synchronization and Wrap-up
related talks
43mAnatomy of an open source voice assistant
36mMyCroft
45mVoice Assistants & IoT Automation
45mBuild your own voice recognition solution with Alexa
26mOpen AI Assistants
35mAgainst all odds, a cloudless smart home
24mHow I ended up automating my curtains and shouting at my laptop We encounter automation every day and may not realise
28mFeeling Lonely? Just Chat with Your House!