DEV.co
Open-Source Observability · ruvnet

RuView

RuView is a WiFi-based sensing platform that detects people, monitors vital signs (breathing and heart rate), and tracks activity through walls using Channel State Information (CSI) from cheap ESP32 sensors—no cameras or wearables required. It integrates with Home Assistant, Apple Home, Google Home, and Alexa, and runs entirely on edge hardware with no cloud dependency.

Source: GitHub — github.com/ruvnet/RuView
78.8k
GitHub stars
10.6k
Forks
Rust
Primary language
MIT
License (OSI-approved)

Key facts

Objective fields from the source. Values we can't verify are shown as “Unknown” rather than guessed.

FieldValue
Repositoryruvnet/RuView
Ownerruvnet
Primary languageRust
LicenseMIT — OSI-approved
Stars78.8k
Forks10.6k
Open issues342
Latest releasev0.8.3-esp32 (2026-06-27)
Last updated2026-07-08
Sourcehttps://github.com/ruvnet/RuView

What RuView is

Written in Rust, RuView captures CSI data from ESP32 mesh networks and processes it through pretrained contrastive encoders (8 KB quantized model on Hugging Face) and spiking neural networks to infer occupancy, vital signs, pose (17-keypoint), and activity. The system supports cryptographic attestation via Ed25519, multi-frequency mesh scanning across 6 WiFi channels, and a 105-cog edge module registry for extensible on-device intelligence.

Quickstart

Get the RuView source

Clone the repository and explore it locally.

terminalbash
git clone https://github.com/ruvnet/RuView.gitcd RuView# follow the project's README for install & configuration

Need it deployed, integrated, or customized instead? DEV.co ships production installs.

Best use cases

Smart home occupancy and wellness monitoring

Detect room presence, breathing rate, heart rate, and sleep quality without cameras—integrates natively with Home Assistant, Matter, Apple Home, Google Home, and Alexa. Useful for elderly care, activity monitoring, and privacy-preserving occupancy automation.

Edge AI research and WiFi sensing prototypes

Open platform with published pretrained models on Hugging Face, self-supervised training pipeline, and 1463 unit tests. Suitable for ML researchers building contactless sensing systems and academic prototypes on commodity hardware.

Industrial and retail spatial intelligence

Multi-person counting, queue detection, customer flow analysis, and facility monitoring via specialized learned counters. Works through walls and in darkness without privacy concerns of video surveillance.

Implementation considerations

  • ESP32 hardware BOM ~$140 for a node plus Cognitum Seed; scale to multiple rooms requires mesh deployment and MQTT broker or Matter bridge infrastructure.
  • Initial ambient calibration takes ~30 seconds per deployment; spiking neural networks adapt locally but may require retraining or fine-tuning in novel RF environments.
  • CSI quality and vital sign accuracy depend on antenna placement, room geometry, and distance to subjects (effective range ~5 m); multipath and through-wall sensing require site validation.
  • Model quantization (8 KB 4-bit) trades accuracy for edge latency; full-precision training on GPU (~2.1 s on RTX 5080) but inference on Pi 5 cold-start ~8.4 ms.
  • Cryptographic attestation (Ed25519 witness chain) is available but integration with existing audit/compliance workflows is not documented; requires review for regulated use.

When to avoid it — and what to weigh

  • Require guaranteed medical-grade accuracy — Vital sign detection is real-time but not clinically validated. The v2 encoder reports 82.3% held-out temporal-triplet accuracy; breathing and heart rate use signal-processing heuristics (bandpass filters) that are environment-dependent and may drift without recalibration.
  • Need production support and SLAs — Project is actively maintained (latest release June 2026) but is open-source without commercial support contracts. Production deployments rely on community issues (342 open), internal expertise, or commercial partnerships (not documented here).
  • Building systems for non-technical users — Requires ESP32 hardware procurement, network configuration, MQTT or Matter bridge setup, and Home Assistant or custom integration. Not a plug-and-play consumer device; technical overhead is moderate to high.
  • Operating in RF-hostile environments — Depends on stable WiFi channel state information. Heavy RF interference, Faraday-shielded spaces, or environments with frequent channel switching may degrade CSI quality and model accuracy.

License & commercial use

MIT License (SPDX: MIT). Permissive open-source license allowing commercial use, modification, and distribution with attribution. No copyleft restrictions.

MIT license permits commercial use without royalty. However, source/dependency licensing of Cognitum Seed, pretrained models on Hugging Face, and any proprietary Cog binaries (aarch64/x86_64 signed binaries on GCS) must be verified independently. Recommend legal review before shipping in commercial products.

DEV.co evaluation signals

Editorial assessment — not user reviews. Directional, with an explicit confidence level.

SignalAssessment
MaintenanceActive
DocumentationAdequate
License clarityClear
Deployment complexityModerate
DEV.co fitGood
Assessment confidenceHigh
Security considerations

No claims made about security posture in provided data. CSI-based sensing is inherently passive (no camera/microphone). Cryptographic attestation via Ed25519 witness chain is mentioned but not detailed; recommend threat modeling for your RF environment. MQTT and Matter integrations should use standard security best practices (TLS, authentication). No penetration test data or vulnerability disclosure policy provided.

Alternatives to consider

mmWave radar (Texas Instruments, Infineon, Qorvo)

Purpose-built for occupancy and vital sign detection; higher accuracy and range (~10 m) but higher cost (~$50–200/node) and more power consumption than WiFi CSI.

Ultra-wideband (UWB) localization (Decawave, Qorvo DW1000)

Sub-meter positioning and occupancy with lower latency; not contactless vital signs but cleaner RF fingerprinting in cluttered environments.

Thermal imaging + edge vision (FLIR Lepton, MLX90640)

Privacy-preserving occupancy and activity via heat signature; no through-wall capability, higher BOM (~$80–150/camera), but deterministic and regulation-friendly in some verticals.

Software development agency

Build on RuView with DEV.co software developers

RuView offers privacy-preserving occupancy and vital sign monitoring without cameras. Assess fit for your RF environment, integration needs, and accuracy requirements. Start with a single ESP32 node (~$9–50) and Home Assistant or Matter bridge.

Talk to DEV.co

Related open-source tools

Surfaced by semantic similarity across the DEV.co open-source index.

Related on DEV.co

Explore the category and the services that help you build with it.

RuView FAQ

Does RuView work through concrete and metal walls?
Yes, CSI-based sensing works through walls within ~5 m (signal-dependent). Multipath modeling and Fresnel-zone geometry allow through-wall detection, but accuracy degrades with distance and RF obstruction. Recommend site validation.
What is the typical latency for presence detection and vital signs?
Presence detection: < 1 ms inference after ~30 s ambient calibration. Breathing rate (0.1–0.5 Hz): real-time, 6–30 BPM range. Heart rate (0.8–2.0 Hz): real-time, 40–120 BPM range. Cold-start pose inference on Pi 5: ~8.4 ms.
Can I train my own models or do I have to use the pretrained weights?
Yes. Pretraining pipeline (60K frames, self-supervised contrastive encoder) is documented. Pose estimation can be retrained in 2.1 s on RTX 5080. World model (OccWorld) supports fine-tuning via `occworld_retrain.py`. Models are published on Hugging Face for reproducibility.
Is RuView suitable for regulatory compliance (HIPAA, GDPR, etc.)?
Unknown from provided data. CSI-based sensing avoids video/audio capture (privacy advantage). Cryptographic attestation and edge processing support compliance, but no documented audit trail, data retention policy, or regulatory certification is provided. Recommend legal/compliance review.

Software development & web development with DEV.co

From first prototype to production, DEV.co delivers software development services around tools like RuView. Our software development agency staffs experienced software developers and web developers for custom software development, web development, integrations, and ongoing support across open-source observability and beyond.

Evaluate RuView for Your Smart Home or IoT Project

RuView offers privacy-preserving occupancy and vital sign monitoring without cameras. Assess fit for your RF environment, integration needs, and accuracy requirements. Start with a single ESP32 node (~$9–50) and Home Assistant or Matter bridge.