← The AI Hype Audit — all 645 verdicts

PARTLY

Partly true: RuView is a real open-source Wi-Fi sensing project with presence and breathing detection, but it needs special sensor hardware in or near the room, it is not new, and its own README says the pose tracking is still weak

Facebook · Nov 2026

The claim'Someone just open sourced software that lets you see through walls using your own Wi-Fi signals', called RuView, on GitHub. It turns everyday Wi-Fi into low-cost radar, uses AI on channel state information to make real-time 3D maps of where you are and what you are doing, and can read your breathing and heart rate 'from anywhere'. 'Would you trust this technology inside your house without you knowing?'

RuView exists: github.com/ruvnet/RuView, MIT licensed, about 104,000 stars on Nov 15, 2026. It builds on real research; Carnegie Mellon's 'DensePose from WiFi' work showed that Wi-Fi channel state information (CSI) can estimate human pose. The project says it detects presence through walls, breathing and heart rate, falls and activity. Three things in the reel are overstated. First, 'your own Wi-Fi' is not enough. The README says presence, vitals and through-wall sensing require CSI from an ESP32-S3 board (about 9 dollars) or a research network card; ordinary laptops only get basic signal-strength presence, and the Docker demo runs on simulated data. Second, 'from anywhere' is wrong; the project's own through-wall range is about 5 meters, so someone has to place sensors near the space. Third, 'real-time 3D mapping of what you are doing' is ahead of the code. The README's honesty table says the on-device 17-point pose model scores 3% at PCK@20 and its runtime path is still a stub, and an earlier '100% presence' figure was retracted. It is not 'just published' either; the repo dates to mid-2025. The privacy question is fair, but the realistic risk is a device planted nearby, not your router turning on you. No independent test of RuView turned up.

What holds up

  • Pulled ruvnet/RuView from the GitHub API: about 104,000 stars, MIT, created June 2025, active Nov 2026.
  • Read the README: CSI hardware (ESP32-S3 or research NIC) required for presence, vitals and through-wall sensing; Docker demo uses simulated data.
  • README honesty table: presence encoder 82.3% held-out accuracy (old 100% figure retracted); on-device pose model PCK@20 of 3.0% and a runtime stub.
  • Project states through-wall range of about 5 meters, signal dependent.
  • One search found no independent third-party evaluation of RuView, only mirrors and the project's own pages.

What doesn’t

  • 'Your own Wi-Fi' omits the required CSI sensor hardware.
  • 'From anywhere' contradicts the project's ~5 m range.
  • 3D pose tracking is first-cut by the project's own numbers.
  • Not new: repo dates to mid-2025.

The catch

The sensing is real and the privacy question is fair, but it takes planted hardware close to the room, and the pose tracking shown is not what the code reliably does yet.

How to actually do it

  • To evaluate it, read the 'Model weights: what's real, what's not' section of the RuView README before anything else.
  • Try the Docker demo first (simulated data), then a single ESP32-S3 node in a room you control for presence and breathing.
  • For elderly-care or presence use, compare it against commercial mmWave presence sensors, which are certified and supported.
  • For privacy, check your own home and office for unknown small powered boards near walls; a router alone does not do this.

Wi-Fi sensing is a real and growing field, and RuView is a serious hobbyist entry point. The reel's 'anyone, anywhere, through your router' framing is the overreach.

Confidence
Medium
Posted by
a tech-commentary creator (about 717,000 plays); reel sent to Buddy Sun Nov 8, 2026, 7:13 AM

See the original claim →

We test hype for free. We build the real thing for a living.

Thirty minutes, no pitch — and you'll leave with something useful either way.

Book a call with Todd or start with the free Business Checkup →

The Verdict Weekly

Three verdicts every Friday. Free forever, unsubscribe anytime, no spam — that would be ironic.

© Schreier Group · schreiergroup.com · See a wild AI claim? Drop it here and we'll test it.