For years, the smart home has struggled with a fundamental conflict: to be truly aware of you, it needs to watch you. This has meant a choice between privacy-invading cameras or crude, unreliable motion sensors. RuView, a massively popular Rust-based project, proposes a third way that feels like it’s pulled from the future: see through walls with WiFi.
It’s not an exaggeration. RuView turns your entire space into a sensor by listening to how human bodies disturb the WiFi signals already bouncing around. By deploying a mesh of inexpensive ($9) ESP32 boards, the system can detect presence, count people, track movement, and even measure breathing and heart rates without contact. It does all of this locally, on edge hardware, with no cameras, no wearables, and no data sent to the cloud. It’s a profound shift in what’s possible for ambient computing.
How It Works: Physics as a Sensor
The core technology is Channel State Information (CSI), a dataset that describes how a radio signal travels from a transmitter to a receiver. Every time you move, or even breathe, you subtly alter the path of these signals. RuView captures these minute changes and feeds them into a compact AI model. The project uses spiking neural networks that can adapt to a new environment in under 30 seconds, learning the unique radio-frequency fingerprint of your space.
What’s particularly clever is its use of surrounding WiFi networks. The system performs multi-frequency scans across six channels, effectively using your neighbors' routers as additional 'illuminators' for its passive radar. Every measurement is then cryptographically signed in a witness chain, creating a verifiable, privacy-first record of events.
Features That Redefine 'Presence Detection'
RuView's feature set goes far beyond a simple binary occupied or unoccupied state. It integrates seamlessly with all major smart home ecosystems—Home Assistant, Apple Home, Google Home, and Alexa—exposing a staggering 21 data points (entities) per sensor node. These include raw signals but, more importantly, ten inferred semantic states. You can build automations based on triggers like someone-sleeping, possible-distress, elderly-inactivity-anomaly, or fall-risk-elevated.
This is where the project shines. It’s not just about knowing if someone is in a room; it’s about understanding the context of their presence. The system offers contactless vital sign monitoring, sleep stage classification, and even camera-free human pose estimation, which can identify 17 body keypoints from radio waves alone. The project is also refreshingly honest about its model's performance, publicly retracting a previous marketing claim and citing a more robust temporal-triplet accuracy of 82.3% for its v2 encoder.
A Platform for Power Users
RuView is more than a tool; it's a full-fledged platform. It includes a complete workflow for capturing CSI data, training your own models, and deploying them. For developers, the most impressive feature might be the RuView MetaHarness. This is a command-line tool that acts as an AI operator for the system. It can provide source-cited guidance, run verification checks, and even use AI agents like Claude Code to help you debug your setup. This is a novel approach to managing the complexity of a sophisticated hardware/software system and demonstrates a deep commitment to user experience.
The project also has forward-looking goals like a "Unified RF world model" to fuse data from WiFi, radar, and UWB, and a "Governed evidence" framework to attach privacy policies directly to the data. These show a level of ambition and thoughtfulness rarely seen in open-source hardware projects.
Rough Edges and Reality Check
Despite the polish, this is not a consumer appliance. The initial setup requires sourcing and flashing firmware to ESP32 boards and deploying a central server on a device like a Raspberry Pi. The documentation also heavily references "Cognitum Seed," which appears to be a related commercial hardware product from the same organization, potentially acting as a preferred, performance-optimized hub. While the system is fully open-source, users should be aware of this commercial ecosystem.
With 88,327 stars, the project has immense interest, but also 434 open issues. This is a healthy ratio for a project of this scale, but it indicates that users will likely encounter bugs or edge cases. Features like the unified world model are still described as "synthetic until real-data validation," so it’s important to distinguish the project's current, impressive reality from its even more ambitious roadmap.
The Verdict
RuView is a landmark project that delivers a powerful, privacy-centric vision for the future of the smart home. It replaces cameras and wearables with clever physics and open-source AI. The barrier to entry is higher than a simple smart plug, but the payoff is an unparalleled understanding of your living space. For anyone serious about home automation, IoT research, or building truly intelligent environments, RuView is an essential tool that is both practical today and pointing the way to the future.