An independent explainer for ruvnet's RuView — built to help you actually implement it.

source github.com/ruvnet/RuView

RuView
WiFi that notices you

Your WiFi can already tell if you're breathing

RuView turns the WiFi signal already filling your rooms into a way to sense the people in them. It works through walls, in the dark, with no camera and nothing to wear. The router notices you; it doesn't watch you.

A single sensor chip, about $9, listens to your WiFi and reports back what it hears: someone's home, someone's moving, someone's breathing.

An independent explainer for rUv (ruvnet)'s RuView — built to take you from "never seen it" to "ready to implement".

Primary use Sense people through WiFi, no camera or wearable.Sensor hardware A $9 chip that listens to WiFi.Ecosystem npm and Rust, plus a Python client.
01

Cameras watch you. Wearables get forgotten.

What problem does this actually solve for me?

Most ways of checking on someone cost you something.

A camera sees a face, every time, and someone has to store that footage somewhere. It can leak, get subpoenaed, or just make people feel spied on in their own home.

A wearable only helps if it's actually being worn. Batteries die. Bands get left on the nightstand. The one moment someone falls might be the one moment their device is off.

So the real question is harder than 'add a sensor.' It's this: can you know someone is safe without recording them, and without asking them to remember anything at all?

The problem
02

RuView: a WiFi signal that quietly keeps watch

Okay, so what is this thing?

RuView reads the WiFi signal already in a room and turns it into information about the people inside.

It detects presence, counts people, and tracks when they come and go. It measures breathing and heart rate, and it can flag a fall. All of this works through walls and in complete darkness, because it was never watching light in the first place.

There's no lens pointed at anyone, no microphone, and nothing to strap on. The same signal that already carries your WiFi is doing the sensing.

The big idea
03

The clever move: your body is already talking to the WiFi

What's the clever trick that makes it work?

Every WiFi router constantly sends out radio waves, and those waves bounce around the room before reaching your phone or laptop.

When you walk through that space, the waves bounce differently. Even when you just sit still and breathe, your chest rising and falling changes the bounce by a tiny, measurable amount.

RuView doesn't add anything new to the room. It just teaches a small, cheap chip to notice the wobble that was already there, and to translate that wobble into presence, movement, and a heartbeat.

The aha

Your WiFi has been recording how you move the whole time — RuView just finally reads it.

04

One level under the hood

How does it actually do that?

Here's the mechanism, one step more technical.

The wobble has a real name: Channel State Information, or CSI. It's basically a fingerprint of how the wave bounced around the room at that instant. A small chip called an ESP32, real hardware costing about $9, captures that fingerprint many times a second.

A model, trained on real recordings of people moving, breathing, and sitting still, turns each fingerprint into an estimate: a body position, a breathing rate, a heart rate. The project is upfront about maturity here. Some checkpoints are validated against real test data. Others are still experimental. The repo labels which is which, instead of blurring the line.

On top of core sensing sit small add-on programs called 'edge modules,' or 'cogs.' Each one runs directly on the sensor chip, with no internet connection required. Each does one specific job — noticing a cough, a glass break, or an overloaded elevator.

Architecture
Architecture — modules, components and how they depend on each other.
Data flow
Data flow — how a request moves through the system at runtime.
05

What people actually build with it

Who uses this, and for what?

Three different people reach for RuView for three different reasons.

In the real world
06

Try it in five minutes, no hardware required

How do I try it myself?

The fastest path uses simulated data, so you can see what it does before buying anything.

docker pull ruvnet/wifi-densepose:latest && docker run -p 3000:3000 ruvnet/wifi-densepose:latest
  1. Install Docker This is the only prerequisite for the simulated path — no ESP32 chip needed yet.
  2. Pull the image Run docker pull ruvnet/wifi-densepose:latest to download the ready-made RuView server.
  3. Run it Run docker run -p 3000:3000 ruvnet/wifi-densepose:latest. This starts a sensing server using simulated data.
  4. Open the dashboard Visit http://localhost:3000 in your browser and watch simulated presence, breathing, and movement data update live.
  5. Check your setup and get guidance Run npx @ruvnet/ruview@0.4.0 doctor to check the local setup, then npx @ruvnet/ruview@0.4.0 guidance --topic sensing --query "model loading" for source-cited help.
  6. When you're ready for the real thing Get an ESP32-S3 board, about $9, flash it, and provision it onto your WiFi using the firmware scripts in the repo — then point RuView at real radio data instead of simulated data.
07

The knowledge pack

Does my AI get it too?

RuView's own documentation and code have been read, chunked, and indexed so both people and AI tools can search it, and check claims against real sources instead of guessing.

# RuView-knowledge-pack.zip for-ai/ # wire this into your agent RuView-kb.rvf # 384-dim vector brain (semantic search) RuView-kb.passages.jsonl # full passage text (search returns TEXT) RuView-symbols.json # exact public API RuView-dep-graph.json # what depends on what RuView-entrypoints.json # build / test / run commands ask-kb.mjs · kb-mcp-server.mjs # CLI + MCP search server for-humans/ # read first RuView-primer.md # the human orientation
Download the knowledge packRVF vector KB + MCP server — drop it into your own agent.
Give your AI the same understandingRuView-knowledge-pack.zip