RuView is a WiFi sensing system that turns ordinary wireless signals into presence detection, movement tracking and contactless breathing and heart-rate readings. It is written mainly in Rust, published under the MIT license, and aimed at people who build their own smart home: Home Assistant users, ESP32 tinkerers, and developers who want room-level awareness without putting a camera in a bedroom or asking anyone to wear a device.
What it does
The project's claim is simple: the WiFi already bouncing around your rooms carries enough information to tell whether someone is there, where they are moving, and how their chest is rising and falling. RuView reads those reflections and turns them into signals a home automation system can act on.
From the materials, that covers:
- presence detection — whether a room is occupied, including through walls and in the dark
- movement and position tracking as people walk around
- vital signs, specifically breathing and heart rate, measured without contact
- continuous room monitoring, published live rather than in batches
The other half of the pitch is integration. RuView speaks to Home Assistant through an MQTT publisher the project calls HA-DISCO, presents itself to Apple Home and HomePod as a discoverable HAP-1.1 bridge, and reaches Google Home and Amazon Alexa either through the same Home Assistant bridge or as a Matter endpoint. The README notes that Siri, Google Assistant and Alexa can then report presence and vitals by room without any custom voice skills being written. It runs offline: no cloud account, no internet connection required.
How it works
The sensing side runs on cheap hardware. The video script describes a nine-dollar ESP32 chip reading WiFi reflections, with a small model running directly on the chip that fits in eight kilobytes and produces results immediately. That is the whole trick the project leans on — no camera, no radar module, no wearable tag, just commodity radio hardware interpreting how signals change as a body moves through them.
Around that sits a Rust codebase, which is where the processing, the bridges and the publishing live. The repository's GitHub topics point at the rest of the shape: firmware and ESP32 for the sensor end, pose-estimation and densepose for the model lineage, home-assistant, home-automation and iot for the output end, and typescript, react and npm for the JavaScript-side pieces. The docs directory carries integration guides and architecture decision records, including one covering how RuView is exposed to Home Assistant over Matter.
Getting started
The lowest-friction path described in the README is an existing Home Assistant install: run RuView with a single --mqtt flag and it announces itself to Home Assistant through discovery. From there, pairing into Apple Home, Google Home, Alexa or SmartThings follows the documented bridge and Matter routes rather than requiring separate integration code.
You will need hardware. Nothing in the project is a pure software install — the sensing depends on an ESP32 board running the project's firmware, so plan on flashing a chip and placing it in the room you want covered. Read the integration document for your ecosystem first; the repository keeps separate guides for the Home Assistant MQTT path and the Apple Home and HomePod bridge, and they describe different setups.
When to use it / when not
This is a good fit when the room is exactly where a camera does not belong. Bedrooms, bathrooms and shared living spaces are the obvious cases, and the offline design means the data does not leave the house. It is also attractive if you already run Home Assistant and want occupancy that is based on a person being present rather than on a phone being connected to the network.
It is a poor fit if you need to know who is in the room rather than that someone is, or if you need a visual record of what happened — this produces signals, not footage. It is also a poor fit if flashing and positioning hardware is not something you want to do.
Be clear-eyed about maturity, too. The repository was created in June 2025 and is still being pushed to, and the materials here contain no accuracy figures, no benchmark and no evaluation of how the vital-sign readings behave across rooms, distances or body positions. Breathing and heart rate measured from radio reflections is a hard problem, and nothing in the project's own description should be read as a medical claim. Treat the numbers it produces as automation inputs to validate yourself, not as measurements to rely on.
RuView deserves attention from home automation builders who have hit the limits of motion sensors and door contacts, and from developers curious about running perception models on nine-dollar microcontrollers. The stack it proposes — cheap radio hardware, a tiny local model, a Rust core, and first-class bridges into every major smart home ecosystem — is unusually complete for a project of this kind, and its 93,000-plus stars suggest the idea has found an audience. Go in expecting to experiment, place sensors, and calibrate against your own rooms rather than to install something finished.