---
title: WaveGuard - Physics-Based Anomaly Detection
emoji: "\U0001F30A"
colorFrom: blue
colorTo: purple
sdk: gradio
sdk_version: "5.29.1"
app_file: app.py
pinned: true
license: mit
tags:
  - anomaly-detection
  - time-series
  - physics
  - waveguard
  - zero-training
short_description: Zero-training anomaly detection using wave equation physics
---

# WaveGuard: Physics-Based Anomaly Detection

**Zero-training anomaly detection powered by wave equation physics.**
No neural networks. No gradient descent. No hyperparameter tuning. **Deterministic.**

Data enters a physics simulation where normal behavior creates smooth wave propagation.
Anomalies disrupt the simulation in measurable ways -- catching threats that statistical
methods miss.

**Ranked #1 on all 12 benchmark datasets** vs. IsolationForest, LOF, and OneClassSVM.

## Demo

Click any scenario button to see pre-computed detection results from the actual WaveGuard engine:

- **Time Series** -- spikes, level shifts, flatlines (4/4 caught, 0% FP)
- **Financial Fraud** -- overseas transactions, structuring, card testing (4/4 caught, 0% FP)
- **Process Health** -- memory leaks, CPU saturation, crashes (4/4 caught, 1 marginal FP)
- **Network Intrusion** -- SYN floods, port scans, blackouts (3/3 caught, noisy FP)

## Use WaveGuard on Your Data

| Method | Link |
|--------|------|
| **RapidAPI** (hosted) | [Subscribe here](https://rapidapi.com/gpartin/api/waveguard) -- Free tier available |
| **Python SDK** | `pip install WaveGuardClient` |
| **MCP Server** | [Glama](https://glama.ai/mcp/connectors/com.emergentphysicslab/waveguard) -- search "waveguard" |
| **PyPI** | [pypi.org/project/WaveGuardClient](https://pypi.org/project/WaveGuardClient/) |
| **GitHub** | [github.com/gpartin/WaveGuardClient](https://github.com/gpartin/WaveGuardClient) |

## How It Works

1. **Encode** -- data mapped into a simulation environment
2. **Evolve** -- wave equations adapt to your normal patterns
3. **Lock** -- evolved state frozen as the reference model
4. **Test** -- new data enters the simulation; anomalies create measurable disruptions
5. **Score** -- disruption magnitude determines the anomaly score with per-feature breakdown

## Key Properties

- **2-15 training samples** (vs hundreds for ML methods)
- **Zero hyperparameters** (sensitivity is the only knob)
- **Deterministic** (same input always gives same output)
- **Explainable** (per-feature scores show what triggered detection)
- **CPU-friendly** (GPU optional, not required)

## Pricing (RapidAPI)

| Plan | Price | Rate Limit |
|------|-------|-----------|
| Free | $0 | 10 req/month |
| Pro | $0.005/req | 60/min |
| Ultra | $0.003/req | 300/min |
| Mega | $0.001/req | 1000/min |

**WaveGuard v3.3.0** by [Emergent Physics Lab](https://huggingface.co/emergentphysicslab)