RuView

RuView turns the WiFi router already in your building into a through-wall sensing system. By analyzing Channel State Information from commodity WiFi signals, it detects human presence, movement, breathing rate, and body posture without cameras, wearables, or any line of sight. The entire system runs on an ESP32-S3 microcontroller that costs about $9, with no cloud connection and no internet required.

The underlying technique comes from CMU’s 2023 DensePose From WiFi research, which proved that WiFi signal distortions caused by the human body contain enough data to reconstruct 24 body regions simultaneously. RuView takes that lab demonstration and packages it into a deployable edge system. The AI model fits in 55KB of memory and adapts to a new environment in under 30 seconds. The primary runtime is written in Rust, delivering an 810x speedup over the original Python implementation, with the Python version retained as a legacy API.

Practical applications include eldercare fall detection without privacy-invasive cameras, smart home presence detection that works through walls, and occupancy monitoring for offices or rental properties. A HomeKit bridge feeds detections directly into Apple Home automations.

One honest caveat: the project comes from a solo developer with a reputation for shipping large numbers of AI-assisted repositories at speed. With 344 branches and an aggressive commit pace, code quality and long-term maintenance are worth evaluating before building any production dependency on it. For hobbyists and indie hackers, the hardware cost is negligible and there is nothing else in open source doing what this does at this price point.