Technical deep dive
How the bridge works under the hood, and why it's built the way it is. None of this is needed to run it; reach for it when you want to understand a decision, extend the project, or debug something the Guide doesn't cover.
How the bridge fits together
- Architecture: the single-process design, its consumers, and the poller state machine.
- Configuration: the full
.envreference, settings precedence, and the schedule syntax. - Dashboard internals: the estimate maths, the trends chart, and the palette rules.
Talking to the gateway
- Register map: every register the poller reads, with types and units.
- Devices & sources: how it discovers multiple inverters, solar strings, and Smart Port loads, and what it doesn't read yet.
- Control registers: the writable surface the bridge deliberately leaves alone, and why.
Interfaces
- HTTP API: every endpoint, the snapshot schema, and the history queries.
- Apple Home mapping: why power becomes a temperature sensor.
- Google Home fulfillment: the intents, the trait mapping, and the stub OAuth.
- Cloudflare Tunnel & Access: publishing the bridge over HTTPS, gated at the edge.
Subsystems
- Energy tariffs: the cost engine and what it can't model.
- Alerts engine: the trigger catalogue, the edge state machine, and its delivery channels.
- Security model: exactly what's open, what the passcode locks, and the recovery path.
Working on it
- Local development: running it locally, the scripts, verifying the gateway, and how the README screenshots and walkthrough are captured.
Also on GitHub
This deep dive mirrors docs/DEEP_DIVE.md in the repository, reorganised for the web.