NINA Mobile for Home Assistant

Unofficial Home Assistant custom integration that makes NINA warnings follow either a GPS-based device_tracker or a fixed Home Assistant zone.

Compatibility: Home Assistant 2025.5.0 or newer (including 2026.6.x). The minimum is set by the current pynina==1.0.2 dependency, which requires Python 3.13.2 or newer.

Important: This project is not an official BBK/NINA integration and is not affiliated with the Bundesamt für Bevölkerungsschutz und Katastrophenhilfe.

What it does

NINA Mobile combines three pieces:

1. A Home Assistant device_tracker or zone provides latitude and longitude.
2. The public BKG VG250 WFS is used to determine the current German district (Kreis) and its ARS.
3. The NINA data source is queried through pynina, using the same current library version as the Home Assistant Core NINA integration at the time of this release.

With a device_tracker, the result behaves like a moving NINA region for a camper, car or other mobile Home Assistant installation. With a zone, NINA Mobile monitors the fixed location of that zone, for example zone.work.

Update behaviour

• NINA warnings are polled every 5 minutes.
• The coordinator uses always_update=False: if region and warning-slot contents are unchanged, Home Assistant entities are not written again just because a poll happened.
• The integration listens to changes of the selected location source independently of that interval. For a device_tracker, this follows GPS movement; for a zone, it also notices if the zone coordinates are edited.
• After the current district polygon has been loaded, GPS movement inside that polygon causes no region lookup and no warning refresh.
• When the GPS point leaves the cached district polygon, the new district is resolved and NINA is refreshed immediately.
• NINA Mobile does not send notifications itself.

Stable warning slots

Warning entities use fixed slots such as:

• binary_sensor.nina_mobile_warnung_1
• binary_sensor.nina_mobile_warnung_2
• ...

A warning ID stays in the same slot for as long as possible. A normal 5-minute refresh therefore does not deliberately reshuffle existing warnings. A new warning is placed into a free slot; a removed warning frees its slot.

This makes the integration easier to consume from automations and other integrations.

Entities

For every configured slot, NINA Mobile creates one safety binary sensor containing the complete warning as attributes, including:

• warning_id and the NINA-compatible alias id
• headline
• description
• sender
• severity
• recommended_actions
• affected_areas
• more_info_url and the NINA-compatible alias web
• sent
• start
• expires
• ars
• region

It also creates diagnostic sensors similar to the standard NINA integration for headline, sender, severity, affected areas, URL and timestamps.

A diagnostic Aktuelle Warnregion sensor shows the district selected from the configured location source and exposes the ARS, coordinates and source entity as attributes.

Location source requirements

The selected entity can be either:

• a GPS-based device_tracker, or
• a Home Assistant zone such as zone.home or zone.work.

Both must expose these attributes:

latitude: 50.1234
longitude: 7.1234

A GPS source can be an ESP32, Raspberry Pi, Home Assistant Companion App, MQTT tracker or anything else that ultimately creates such a device_tracker. NINA Mobile does not control its GPS update interval, so the tracker can implement its own adaptive update logic based on movement or speed.

A zone uses the fixed center coordinates configured in Home Assistant. This is useful for a second NINA Mobile instance for a place such as work, family or a campsite.

Installation through HACS as a custom repository

1. Create a public GitHub repository, for example nina-mobile.
2. Upload the complete contents of this repository.
3. In HACS, add the GitHub repository as a custom repository of type Integration.
4. Install NINA Mobile.
5. Restart Home Assistant.
6. Go to Settings → Devices & services → Add integration → NINA Mobile.
7. Select a GPS device_tracker or a zone and the desired number of warning slots.

Recommended GitHub repository metadata:

• Description: GPS-following NINA warning integration for Home Assistant
• Topics: home-assistant, hacs, nina, bbk, warnings, gps, device-tracker, germany
• Issues: enabled

Manual installation

Copy:

custom_components/nina_mobile/

into:

/config/custom_components/nina_mobile/

and restart Home Assistant.

Behaviour when coordinates are unavailable

If the selected source temporarily stops providing coordinates, NINA Mobile keeps the last successfully resolved district and continues polling warnings for it. It does not immediately clear an active warning simply because coordinates disappeared for a moment.

If a valid location is resolved outside Germany, the region is cleared and the warning slots turn off.

Known limitations in v0.1.2

• The integration follows the district/ARS model used by the Home Assistant NINA integration. It does not yet test the GPS point against each individual NINA warning polygon.
• Near an administrative boundary, inaccurate GPS positions can make the selected district switch back and forth. Boundary hysteresis can be added later.
• The public BKG VG250 WFS product page currently lists its administrative data as 01.01.2025, while a newer downloadable VG250 dataset exists. This first version deliberately uses the WFS because it avoids bundling a large Germany-wide geometry file.
• This is an initial implementation and should be tested with real Home Assistant/NINA data before being treated as safety-critical.

Data sources and attribution

NINA / BBK

Warning data is retrieved from the public NINA warning infrastructure (warnung.bund.de) through the pynina Python package.

BKG administrative districts

District geometry and ARS information are requested from the public BKG VG250 WFS:

https://sgx.geodatenzentrum.de/wfs_vg250

Attribution for the currently published WFS dataset:

© GeoBasis-DE / BKG 2025, dl-de/by-2-0. Data sources: BKG / German surveying authorities as documented by the BKG.

License

The integration source code is released under the MIT License. External warning and geodata remain subject to the terms of their respective providers.