What's new in Kiosk Satellite v0.15.0-beta
This release adds a complete low-latency WebRTC/WHEP camera experience designed for wall tablets, with native controls on the device, matching administration from a browser, and dedicated Home Assistant entities.
New Features
-
WebRTC camera views. Import streams directly from Go2RTC or add cameras manually using a WHEP endpoint, then combine up to four cameras into a named view. Views automatically arrange themselves for the display orientation and aspect ratio, with an option to show camera names. Go2RTC connections support basic authentication, self-signed certificates, missing-stream detection, and a separate high-resolution stream for fullscreen focus.
-
Fast camera navigation built for touch. Tap a camera in a multi-camera view to focus it while the other streams disconnect. Press Back, use the back swipe gesture, or double tap to return to the grid. The same controls close the full view when no camera is focused.
-
Camera controls in Home Assistant. MQTT discovery now creates a button for every camera view, a Close camera view button, and an Active camera view sensor. Each view uses a stable internal ID, so renaming it does not replace its Home Assistant entity.
-
Complete camera management on the device and in Remote Admin. Add and edit Go2RTC servers, import streams, configure manual cameras, order cameras into views, choose whether names are shown, preview views, and remove cameras or views from either interface.
-
Camera views cooperate with voice interactions. A voice interaction temporarily closes the camera view so Voice Satellite remains visible, then restores the same view and focused camera when the interaction finishes. Explicitly closing or changing the view cancels the return.
-
Screensaver-aware camera viewing. Opening a camera view immediately dismisses an active screensaver, and the screensaver remains suppressed until the camera view closes.
-
Optimized for low-memory tablets. The camera player exists only while a view is visible, one WebView owns all peer connections, camera audio is never negotiated, and focusing a camera closes the hidden streams to reduce decoder, CPU, and memory pressure.