S2JYarbo Home Assistant

!!!!Due to the recent security issue with all Yarbo's, the way we interface with the Yarbo is in flux. Until this has been tied down, Expect this Module to not work with your Yarbo. We'll try and keep up to date on this evolving issue and fix things as and when then become available.!!!!

Home Assistant custom integration workspace for Yarbo devices using MQTT, published under the s2jyarbo domain to avoid conflicts with other Yarbo integrations.

This integration allows a user to view the current map of a YArbo, start and stop plans, Edit and Add No Go Zones and Pathways. More to come.

This repository contains:

• the s2jyarbo custom integration in custom_components/s2jyarbo
• a custom S2JYarbo topics sidebar panel
• a custom S2JYarbo overview card for the Home Assistant dashboard

The integration is currently MQTT-first. It subscribes to snowbot/<serial>/#, captures topic samples, builds device summaries from device/DeviceMSG, device/heart_beat, and device/data_feedback, and exposes those details through Home Assistant entities and custom UI.

See CHANGELOG.md for feature history.

Overview snapshot

Current capabilities

• Config flow with:
  • broker host or IP
  • broker port
  • TLS on or off
  • Yarbo serial number
• Local-push MQTT runtime with automatic reconnect
• Subscription to all device topics under snowbot/<serial>/#
• Topic discovery and sample capture in the S2JYarbo Topics sidebar panel
• Pairing of app/get_* and app/read_* commands with device/data_feedback
• Merged device/DeviceMSG document to capture fields that appear across multiple messages
• Diagnostic sensor entities for MQTT/runtime state
• Device tracker entity from GPS / RTK data
• Overview dashboard card with one widget per selected Yarbo device
• Integrated local map widget inside the overview card with:
  • decoded get_map geometry rendering
  • live Yarbo footprint and heading with fixed GPS offset correction
  • dock geometry and docking guard square rendering
  • plan_feedback overlay rendering with visited vs remaining plan path state
  • plan summary pill with plan name, progress, remaining area, and estimate
  • recharge_feedback return-to-dock route rendering
  • cloud_points_feedback collision/barrier rendering
  • follow mode, zoom, pan, and breadcrumb trail controls
  • in-map GPS/heading overlay
  • variable trail width based on mower motor state
  • magenta reverse trail segments
  • edit mode for pathways and no-go zones after a stored Home Assistant acknowledgement
  • pathway/no-go zone point editing, whole-object move, rotate, and no-go zone sizing controls
  • no-go zone preset/context actions for circle, 1 m square, and 1 m 8-point circle generation
  • unsaved-change protection while editing, including Home Assistant dashboard navigation
• Plan dropdown populated from read_all_plan
• Start percentage slider under the plan selector
• Device commands from the overview card:
  • start plan
  • pause
  • resume
  • stop
  • recharge
  • shutdown
  • restart
  • volume update
  • bulk refresh
• Auto-refresh of stale device data when Last Updated is missing or older than one hour
• Three custom cards to add to your dashboard.

Project layout

.
├── .github/workflows/          # Validation workflows
├── .homeassistant/              # Local Home Assistant config directory
├── custom_components/
│   └── s2jyarbo/                # S2JYarbo custom integration
├── scripts/                     # Local workflow helpers
├── CHANGELOG.md                 # Project change log
├── LICENSE                      # Distribution license
├── docker-compose.yml           # Local Home Assistant instance
├── hacs.json                    # HACS metadata
├── pyproject.toml               # Ruff configuration
└── requirements-dev.txt         # Optional local tooling

Install in Home Assistant

HACS

Once this repository is published and available on GitHub, the recommended install path is HACS.

1. Open HACS -> Integrations
2. Open the menu and choose Custom repositories
3. Add:
   • Repository: https://github.com/steves2j/Yarbo_HomeAssistant
   • Category: Integration
4. Search for S2JYarbo Home Assistant Integration
5. Install it
6. Restart Home Assistant
7. Add the integration from Settings -> Devices & Services

Manual install

1. Copy custom_components/s2jyarbo into your Home Assistant config directory under custom_components/
2. Restart Home Assistant
3. Add the integration from Settings -> Devices & Services

Example target path:

<config>/
└── custom_components/
    └── s2jyarbo/

Release and distribution notes

The repository now includes the minimum scaffolding for a normal HACS-style custom integration release:

• hacs.json
• LICENSE
• GitHub workflow for HACS validation
• GitHub workflow for hassfest

For a clean public release:

1. Push the repository to GitHub
2. Make sure the repository description, topics, and branding are set appropriately on GitHub
3. Bump custom_components/s2jyarbo/manifest.json version
4. Update CHANGELOG.md
5. Create a GitHub release/tag that matches the integration version
6. Verify the GitHub Actions pass

If the GitHub repository URL changes, update:

• custom_components/s2jyarbo/manifest.json
• this README

Home Assistant UI

Devices & Services

Each config entry creates:

• sensor.<name>_mqtt_connection
• sensor.<name>_discovered_mqtt_topics
• device_tracker.<name>_location

The integration can be reconfigured from the S2JYarbo card menu in Settings -> Devices & Services.

S2JYarbo Topics sidebar

The custom admin-only sidebar panel is registered as S2JYarbo Topics.

It shows:

• discovered topics
• latest sample per topic
• command/response pairing for app/get_* and app/read_*
• packet metadata
• merged DeviceMSG document

S2JYarbo dashboard cards

The frontend registers three separate Lovelace cards:

• custom:s2jyarbo-control-card
• custom:s2jyarbo-map-card
• custom:s2jyarbo-advanced-card

Recommended:

• add the three cards to a normal Lovelace dashboard
• configure each card for the same S2JYarbo device when you want a complete dashboard view

If you are using Home Assistant 2026.4 or newer, be aware that new installs use the new auto-generated Home/Overview dashboard by default. That view only exposes Add entity and does not behave like a normal Lovelace dashboard for custom cards.

In that case, the working path is:

1. Open Settings -> Dashboards
2. Click Add dashboard
3. Create a new dashboard
4. Open that dashboard
5. Click Edit dashboard
6. Click Add card
7. Add the S2JYarbo custom cards you want

If you prefer, you can also create a dashboard from the Overview (legacy) template and add the card there.

Example Lovelace setup:

type: custom:s2jyarbo-control-card
---
type: custom:s2jyarbo-map-card
---
type: custom:s2jyarbo-advanced-card

The control card includes:

• connection state
• serial number
• battery and recharge action
• Wi-Fi signal strength
• satellite count
• plan selection
• start percentage slider
• start / pause / resume / stop controls

The map card includes:

• decoded site geometry from get_map
• follow mode, zoom, pan, reset, breadcrumbs, and live device heading
• fixed GPS offset correction applied to the live Yarbo position
• trail width changes between transit and 550 mm cutting width based on mower motor state
• magenta trail segments for reverse movement
• plan_feedback overlays:
  • visited vs remaining plan path rendering
  • right-side plan summary pill
• recharge_feedback as a cyan dotted return-to-dock route
• cloud_points_feedback as tomato collision/barrier strips
• pathway, memory path, and no-go zone edit mode:
  • full-browser warning acknowledgement stored in Home Assistant
  • point insert, drag, and delete
  • Ctrl-drag whole-object move
  • Shift-scroll rotate
  • Ctrl-scroll no-go zone resize
  • no-go zone context menu actions for ToCircle, addSquare, and addCircle
  • unsaved-change prompts with discard, cancel, and save flows
• background aerial image support:
  • upload a background image
  • Shift-drag to move the image
  • scroll to scale the image
  • Ctrl-scroll to rotate the image around the cursor
  • opacity control
  • placement import/export JSON for backup/restore across Home Assistant instances

The advanced card includes:

• status, battery, firmware, GPS, network, and diagnostic details
• shutdown and restart controls
• volume control
• Wi-Fi details

MQTT behavior

The integration currently expects:

• requests on snowbot/<serial>/app/...
• live state from:
  • device/DeviceMSG
  • device/heart_beat
  • device/data_feedback

Notes:

• device/DeviceMSG payloads may be zlib-compressed JSON
• device/data_feedback is paired back to the originating command topic, preferably using the topic field inside the payload
• some brokers do not echo outbound app/* publishes back to subscribers; the integration does not rely on that echo for core device state updates

Development notes

• Code changes in custom_components/s2jyarbo are mounted directly into the local Home Assistant container.
• Frontend changes usually require a hard browser refresh.
• Backend changes usually require reloading the integration or restarting Home Assistant.
• The custom panel/frontend bundle is served from custom_components/s2jyarbo/panel/.

Known limitations

• This is still a custom integration under active development, not a finished package.
• Some Yarbo MQTT payloads and command shapes are being derived from observed traffic and cached samples.
• Native HA recorder may warn when topic/entity attributes become very large.

Reference docs

• Home Assistant development environment:
  • https://developers.home-assistant.io/docs/development_environment/
• Home Assistant config flows:
  • https://developers.home-assistant.io/docs/core/integration/config_flow/
• Home Assistant custom integration branding:
  • https://developers.home-assistant.io/docs/core/integration/brand_images
• Home Assistant frontend icons:
  • https://www.home-assistant.io/docs/frontend/icons/

AI acknowledgement

There is no pretence. Most of this code was developed using AI. AI is a tool, not a replacement for a developer, and it handles a lot of the boring boilerplate work. It still makes enough mistakes that a human developer is required, as it is not yet at a stage where it can write unsupervised code reliably. Do not hate AI; see it as a tool, the next generation of autocomplete. Autocomplete with benefits.