RuView: $7 ESP32 Turns WiFi Into Through-Wall Spatial Radar
> A router and a thumb-sized dev board—no camera, no wearables—can detect through walls which room you're in, whether you're walking or sitting, your breathing rate, your heart rate, and even whether you experience sleep apnea.
RuView, which hit GitHub trending last week, confronts developers with a counterintuitive physical fact: the WiFi signals from your router are already reflecting off your body. Every breath-induced chest movement and every heartbeat-induced micro-vibration perturbs those 2.4GHz electromagnetic waves. The question isn't whether sensing is possible—it's how to decode the signal.
RuView's answer is surprisingly simple: a $7 ESP32-S3 board flashed with open-source firmware captures WiFi CSI (Channel State Information), feeds it to a small locally-running model, and outputs 17-keypoint pose estimation, breathing rate, heart rate, fall detection, and sleep staging.
Physics: WiFi Signals Are Already "Radar"
The first step to understanding RuView is abandoning the intuition that only cameras can "see."
Radar works by emitting electromagnetic waves, receiving reflections, and extracting target position and velocity from delays and Doppler shifts. WiFi routers emit thousands of pulses per second, and your body—mostly water—reflects part of them. Traditional WiFi chips only care about decoding 0s and 1s, discarding reflections as noise. CSI is the discarded part: it records the amplitude and phase of every subcarrier at every moment—a fine-grained snapshot of how the signal was distorted in space.
- A person walking draws a clear trajectory in the CSI waveform.
- A person sitting still breathing produces periodic micro-fluctuations at 12-20 cycles/min.
- A sleeping person shows, alongside the slow breathing cycle, high-frequency micro-tremors at 60-100 cycles/min corresponding to heart rate—two real peaks visible on the FFT spectrum, not a metaphor.
someone-sleepingpossible-distressroom-activeelderly-inactivity-anomalymeeting-in-progressbathroom-occupiedfall-risk-elevatedbed-exitno-movementmulti-room-transition
RuView connects the full chain: ESP32 captures CSI → preprocessing/denoising → trained model → semantic events ("someone is sleeping," "possible fall," "bathroom occupied").
Engineering: 21 Entities Per Smart Home Node
RuView isn't just a paper—it's a working system. Each ESP32 node outputs 21 entities: 11 raw signals (CSI amplitude, phase, SNR, etc.) plus 10 inferred semantic states:
This is a three-layer "sensing → inference → semantics" architecture, isomorphic to computer vision's "pixels → features → object recognition." The difference: RuView's "pixels" are CSI time series, not RGB.
Privacy: Surveillance Without Lenses
RuView's most discussion-worthy aspect isn't technical—it's how it redefines "surveillance."
The problem with cameras isn't that they capture images; it's that captured footage can be replayed, stolen, or used to train face recognition models. RuView has no lens. Its output—"a person is present, breathing 16/min, posture lying down"—is worthless to face recognition models but enormously valuable for "is an elderly person living alone safe?"
The privacy cost of surveillance depends not on what can be sensed, but on whether sensed signals can be inverted back into personal biometric identity. CSI is a highly semantic abstraction: it tells you "a person is breathing" but not "who that person is." That's fundamentally different from cameras.
CSI isn't perfectly anonymous, though: research has shown gait CSI patterns can identify individuals—everyone's walking rhythm, stride, and weight shifts differ. RuView's documentation explicitly acknowledges this, attaching privacy policies, uncertainty, provenance, and witness records to each sensing event. It's a "governance-first" engineering attitude rather than "ship first, ask later."
MetaHarness: An "Honesty Check" for AI Claims
A less conspicuous but important part of the repo is MetaHarness—an AI operator that guides humans and AI agents through deployment, calibration, training, and validation, and performs "honesty checks" on all sensing claims.
Why? WiFi sensing has a chronic problem: 95% accuracy in papers can drop to 60% in your living room—different router placement, wall materials, reflective clutter, lab-trained models. MetaHarness requires every claim to carry source citations, deterministic verification, and field-data validation. A claim isn't capability; a claim plus evidence is capability. The author built this into the system design rather than leaving it to users.
Cross-Domain Analogy: Electric Fish and Bats
RuView's physics has direct biological parallels. Weakly electric fish navigate and hunt in the murky Amazon by emitting electric pulses and sensing reflected field changes; bats use ultrasonic echolocation. Both build environment models from actively emitted signals plus fine-grained reflection analysis—no lens-based vision.
RuView is the artificial version: emit WiFi, parse CSI reflections, build a spatial model. Its "unified RF world model" aims to unify WiFi CSI, millimeter-wave radar, UWB, and cellular sensing into one scene model—mirroring biological multimodal fusion (owls use vision + hearing; sharks use smell + electroreception). Single modalities have blind spots; fusion is robust.
Why Now
WiFi sensing isn't new—MIT's WiTrack paper (2013) demonstrated through-wall tracking. The field was stuck on three bottlenecks:
1. Hardware: extracting CSI required specialized SDRs or driver hacks. The ESP32-S3 is the first cheap consumer chip that outputs CSI directly. 2. Models: traditional hand-crafted features generalized poorly. Transformers and LoRA make small-dataset training viable. 3. Ecosystem: no standardized semantic output layer. RuView's 21-entity model plus Home Assistant/Matter/HomeKit integration is the first to bring WiFi sensing into standard smart home protocols.
With all three loosening simultaneously, $7 hardware + open firmware + pretrained models + smart home integration moves through-wall sensing from the lab to a purchasable DIY kit.
A Deeper Observation
RuView illustrates a universal principle: the best sensor is one you've already deployed but never used.
WiFi routers have always been emitting and receiving reflections. For decades we used only the "can I decode 0s and 1s" layer, discarding fine-grained CSI as noise. RuView's contribution isn't a new sensing method—it's recovering discarded information. The corollary: many "free sensors" around you are being wasted—Bluetooth beacons, ambient light sensors, microphone noise floors, accelerometer micro-vibrations. RuView provides a reusable engineering template for decoding signals that already exist.
$7, one ESP32, one open-source repo. Watching your breathing through a wall, with no lens. Not science fiction—2026's GitHub trending.
---
Project: https://github.com/ruvnet/RuView Model repo: https://huggingface.co/ruvnet/wifi-densepose-pretrained Author blog: https://ruview.blog/wifi-csi-sensing-open-source