HomeBrain
A local-first home automation and voice platform spanning web, mobile, watch, wall-panel, direct-radio, security, telemetry, and private on-device AI experiences.
01 / The problem
What needs to change
Connected homes often depend on fragmented vendor apps and cloud services, leaving automation, voice, security, telemetry, and operations without one coherent local control layer.
02 / The approach
A focused path through the complexity
HomeBrain centers the system on a self-hosted hub that combines direct device control, local voice and model services, cross-surface experiences, and operational tooling that can be managed from the product itself.
03 / Key capabilities
What the system is designed to do
- Device, scene, workflow, security, and climate control
- Local wake word, speech recognition, language models, and speech output
- Native iPhone and Apple Watch companion experiences
- Direct Zigbee, Z-Wave, Matter, Thread, and legacy integrations
- Managed domains, SSL, dynamic DNS, backups, and platform deployment
04 / Architecture
Technology and system shape
- Node and Express services with a React control surface
- Local model, speech-to-text, and text-to-speech services
- Native iOS, watchOS, remote-listener, and embedded-panel clients
- Event distribution through a local MQTT broker
- Linux and Jetson deployment behind a managed reverse proxy
05 / Results
What the repository demonstrates
- The current platform unifies automation, local AI, multiple device protocols, security, telemetry, notifications, and hub operations.
- Companion apps and dedicated hardware surfaces extend the same control model beyond the browser.
06 / Lessons
What the work teaches
- Local-first AI becomes more valuable when model ownership, device state, and privacy boundaries are designed together.
- A home platform must make operations and recovery as approachable as everyday automation.
07 / Screenshots
Approved visuals
No public screenshots yet
Add only approved screenshots to the project record.