# Wavey > Open-source WiFi CSI sensing on ESP32: device-free occupancy, motion, and presence detection without cameras or wearables. Occupancy, motion, and presence from the WiFi environment. Wavey reads WiFi Channel State Information (CSI) from low-cost ESP32 nodes and turns the way a moving body disturbs WiFi signals into occupancy, motion, and presence information - device-free and camera-free. Project links: - Website: https://waveyhq.dev/ - Console (live CSI visualization): https://console.waveyhq.dev - GitHub (source code): https://github.com/waveyhq - Discord (community): https://discord.gg/sxh9r9UTtW - X (Twitter): https://x.com/nopeJS - Email: mail@waveyhq.dev ## Start here - [How WiFi CSI Sensing Works](https://waveyhq.dev/how-it-works/): The physics of WiFi CSI sensing — OFDM subcarriers, multipath superposition, phase vs amplitude, micro-Doppler from breathing, and why commodity WiFi links can sense a room. - [Getting Started with Wavey](https://waveyhq.dev/getting-started/): Deploy Wavey — ESP32 CSI node architecture, Python pipeline layers, detection task selection, and how to follow development on GitHub. - [WiFi CSI Sensing FAQ](https://waveyhq.dev/faq/): Practical questions about WiFi CSI sensing and Wavey — hardware, privacy, accuracy, through-wall sensing, and how to get started. - [WiFi Sensing Glossary](https://waveyhq.dev/glossary/): Technical definitions for WiFi CSI sensing — OFDM subcarriers, multipath, micro-Doppler, phase sanitization, spectrograms, domain shift, HAR, and more. - [About Wavey](https://waveyhq.dev/about/): Wavey is an open-source WiFi CSI sensing project — subcarrier physics, signal processing pipelines, and deployable detection on commodity ESP32 hardware. ## Section indexes - [WiFi CSI Sensing Use Cases](https://waveyhq.dev/use-cases/): WiFi CSI sensing applications — occupancy, HAR, micro-presence, elder-care, security, automation, and analytics. Each page covers one technical angle. - [WiFi CSI Sensing vs Cameras, Radar & Wearables](https://waveyhq.dev/comparisons/): Technical comparison of WiFi CSI sensing vs cameras, mmWave radar, PIR, and wearables — sensing mechanism, latency, coverage, and deployment cost. - [Posts](https://waveyhq.dev/posts/): Technical essays on WiFi CSI sensing — channel structure, commodity hardware limits, activity classifiers, and baseline engineering from the Wavey project. - [Contact Us](https://waveyhq.dev/contact/): Get in touch with the Wavey team for questions, contributions, or ideas about open-source WiFi CSI sensing. ## Use cases - [Device-Free Occupancy Detection with WiFi CSI](https://waveyhq.dev/use-cases/occupancy-detection/): Binary occupancy from WiFi CSI change-point detection on an empty-room baseline — why it tolerates noise, how it differs from MAC counting, and HVAC control semantics. - [WiFi Motion & Activity Detection](https://waveyhq.dev/use-cases/motion-activity-detection/): Human activity recognition on WiFi CSI — CSI-speed modeling, dual-stream transformers, spectrogram classifiers, and cross-site transfer. - [Presence & Breathing Detection with WiFi CSI](https://waveyhq.dev/use-cases/presence-detection/): Micro-Doppler presence sensing from WiFi CSI — respiration band physics, SNR requirements, still-person detection, and why PIR cannot see periodic chest motion. - [Elder-Care Monitoring with WiFi Sensing](https://waveyhq.dev/use-cases/elder-care-monitoring/): Ambient elder-care sensing with WiFi CSI — anomaly-based fall detection, inactivity monitoring, online adaptation, and non-medical boundaries. - [WiFi Sensing for Security & Intrusion Detection](https://waveyhq.dev/use-cases/security-intrusion-detection/): Zone-level intrusion detection with WiFi CSI — NLOS motion through drywall, darkness-invariant sensing, device-free vs MAC tracking, and what you learn vs what you cannot. - [Smart-Home Automation with WiFi Presence Sensing](https://waveyhq.dev/use-cases/smart-home-automation/): WiFi presence sensing for home automation — presence-hold vs motion-pulse events, debouncing, hysteresis, and why occupancy beats PIR for HVAC and lighting scenes. - [Privacy-First Occupancy Analytics with WiFi CSI](https://waveyhq.dev/use-cases/occupancy-analytics/): Anonymous occupancy analytics from WiFi CSI — temporal aggregation, utilization metrics, zone heatmaps, GDPR-safe sensing vs MAC tracking, and honest headcount limits. ## Comparisons - [WiFi Sensing vs Cameras for Occupancy & Motion](https://waveyhq.dev/comparisons/wifi-sensing-vs-cameras/): WiFi CSI vs cameras — sensing mechanism, data payload, latency, coverage geometry, and when each modality fits occupancy vs identification tasks. - [WiFi CSI Sensing vs mmWave Radar](https://waveyhq.dev/comparisons/wifi-sensing-vs-mmwave-radar/): WiFi CSI vs mmWave radar — RF front-end cost, ranging precision, fine-motion SNR, deployment scale, and when dedicated radar beats commodity WiFi. - [WiFi CSI Sensing vs PIR Motion Sensors](https://waveyhq.dev/comparisons/wifi-sensing-vs-pir/): WiFi CSI vs PIR — infrared motion cone vs radio channel sensing, still-detection mechanism, coverage geometry, and why PIR false-offs happen. - [WiFi CSI Sensing vs Wearables & Phone Tracking](https://waveyhq.dev/comparisons/wifi-sensing-vs-wearables/): WiFi CSI vs wearables and MAC tracking — device-free body sensing vs opt-in device telemetry, coverage completeness, and data model differences. ## Posts - [IEEE 802.11bf and the WiFi Sensing Privacy Gap](https://waveyhq.dev/posts/ieee-80211bf-privacy-gap/): 802.11bf standardizes WiFi sensing but leaves privacy protection vendor-dependent. No mandatory secure-LTF for sensing NDPs, no opt-out mechanism, and BFI remains plaintext — what enterprise buyers and researchers need to know. - [RF Privacy Defenses: Obfuscation, Metasurfaces, and Their Limits](https://waveyhq.dev/posts/rf-privacy-defenses-and-limits/): How to prevent unauthorized WiFi sensing — CSI fuzzing, intelligent reflecting surfaces, BFI obfuscation, and anti-sensing shields. What each defense achieves, what it costs, and why detection and mitigation must work together. - [Detecting Unauthorized RF Sensing: A Layered Approach](https://waveyhq.dev/posts/detecting-unauthorized-rf-sensing/): Five detection layers for unauthorized RF sensing — from emitter inventory to behavioral baselines. What works today on ESP32 mesh, what remains open research, and why passive sniffer detection is the hardest problem. - [WiFi Sensing Attacks: LeakyBeam, BFId, and the BFI Side Channel](https://waveyhq.dev/posts/wifi-sensing-attacks-leakybeam-bfid/): NDSS 2025 and CCS 2025 proved passive WiFi surveillance is practical — occupancy detection at 20m through walls, identity inference across 197 subjects, all from plaintext beamforming feedback on commodity hardware. - [RF Privacy: The Flip Side of WiFi Sensing](https://waveyhq.dev/posts/rf-privacy-the-flip-side-of-wifi-sensing/): WiFi CSI sensing works both ways — the same physics that detects occupancy without cameras also lets an attacker observe a room without consent. What the research says and why RF privacy is an unserved problem. - [Commodity CSI: Capture Limits and Calibration](https://waveyhq.dev/posts/esp32-csi-explained/): What ESP32 CSI actually gives you — sampling rate, antenna count, phase noise floor, ESP-CSI API surface, and when cheap hardware is enough for real sensing. - [Baselines, Drift, and Generalization](https://waveyhq.dev/posts/device-free-occupancy-sensing-explained/): Empty-room CSI fingerprinting, furniture drift, cross-site domain shift, adaptive baselines, and when to re-baseline — the engineering behind reliable occupancy sensing. - [Building an Activity Classifier on CSI](https://waveyhq.dev/posts/wifi-motion-detection-without-camera/): A practical path from variance thresholds to spectrogram CNNs for WiFi HAR — label collection, training, evaluation, and why transfer learning across rooms fails without adaptation. - [The WiFi Channel as a Distributed Sensor](https://waveyhq.dev/posts/introduction-to-wifi-csi/): What one CSI matrix actually looks like — OFDM structure, multipath superposition as a worked example, and why RSSI discards the degrees of freedom that make WiFi sensing possible. ## Tags - [802.11bf](https://waveyhq.dev/tags/802.11bf/): 1 pages - [Attacks](https://waveyhq.dev/tags/attacks/): 1 pages - [Baseline](https://waveyhq.dev/tags/baseline/): 1 pages - [BFI](https://waveyhq.dev/tags/bfi/): 1 pages - [Calibration](https://waveyhq.dev/tags/calibration/): 1 pages - [CSI](https://waveyhq.dev/tags/csi/): 1 pages - [Defense](https://waveyhq.dev/tags/defense/): 1 pages - [Detection](https://waveyhq.dev/tags/detection/): 1 pages - [Domain Shift](https://waveyhq.dev/tags/domain-shift/): 1 pages - [ESP32](https://waveyhq.dev/tags/esp32/): 1 pages - [HAR](https://waveyhq.dev/tags/har/): 1 pages - [Hardware](https://waveyhq.dev/tags/hardware/): 1 pages - [Machine Learning](https://waveyhq.dev/tags/machine-learning/): 1 pages - [Multipath](https://waveyhq.dev/tags/multipath/): 1 pages - [Obfuscation](https://waveyhq.dev/tags/obfuscation/): 1 pages - [Occupancy](https://waveyhq.dev/tags/occupancy/): 1 pages - [OFDM](https://waveyhq.dev/tags/ofdm/): 1 pages - [Privacy](https://waveyhq.dev/tags/privacy/): 1 pages - [Research](https://waveyhq.dev/tags/research/): 5 pages - [RF Privacy](https://waveyhq.dev/tags/rf-privacy/): 2 pages - [Security](https://waveyhq.dev/tags/security/): 4 pages - [Spectrogram](https://waveyhq.dev/tags/spectrogram/): 1 pages - [Standards](https://waveyhq.dev/tags/standards/): 1 pages - [WiFi CSI](https://waveyhq.dev/tags/wifi-csi/): 1 pages ## Categories - [Fundamentals](https://waveyhq.dev/categories/fundamentals/): 2 pages - [Research](https://waveyhq.dev/categories/research/): 5 pages - [Technical](https://waveyhq.dev/categories/technical/): 2 pages ## Key facts - Wavey is open source and developed at https://github.com/waveyhq. - It uses WiFi CSI (Channel State Information), which is far richer than RSSI, captured per OFDM subcarrier. - It runs on the ESP32 family (ESP32, S2, S3, C3, C5, C6, ...), which exposes CSI for sensing. - It is device-free (nothing carried by the person) and camera-free (no images), making it privacy-preserving. - It builds on WiFi-sensing research (e.g. CMU DensePose-from-WiFi) and the IEEE 802.11bf sensing standard.