Physics-Based Anomaly Detection

WaveGuard
Zero-Training
Anomaly Detection

Detect anomalies in any data stream using Klein-Gordon wave physics. No training data. No model fitting. No labeled examples. Just send your data and get a physics fingerprint with anomaly scores in milliseconds.

Live on Modal (GPU)37 endpointsPython SDK on PyPIMCP + x402 ready
F1 0.88
Server Metrics
Zero-training
<500 ms
Response Time
GPU-accelerated
37
API Endpoints
Scan, Fingerprint, Compare
$0.001
Per Scan
x402 micropayment

How It Works

WaveGuard treats your data as a wave signal on a 3D lattice and applies the same physics that governs gravitational waves and quantum fields.

1

Encode to Lattice

Your data is mapped onto a 3D cubic lattice using GOV-01 (Klein-Gordon) encoding. JSON, time-series, tabular — all become wave amplitudes.

2

Evolve with Physics

The lattice evolves under GOV-01 + GOV-02 wave equations on GPU. Normal data produces stable patterns; anomalies create instabilities and resonances.

3

Read the Fingerprint

Statistical analysis of the evolved field (energy, chi-well depth, spectral peaks) produces a physics fingerprint. Mahalanobis distance gives the anomaly score.

Works on Any Data

No domain-specific training. The physics doesn't care what your data means — it detects structural anomalies in the signal itself.

Server Monitoring

CPU spikes, memory leaks, latency anomalies. F1 = 0.882 on production metrics.

Crypto Markets

Pump-and-dump detection, wash trading, abnormal volume patterns.

IoT Sensors

Equipment failure prediction, drift detection, environmental monitoring.

Financial Data

Transaction fraud, market manipulation, unusual trading patterns.

Start in 30 Seconds

One POST. One fingerprint. No setup required.

cURL
curl -X POST \
  "https://api.emergentphysicslab.com/waveguard/v1/scan" \
  -H "Content-Type: application/json" \
  -d '{
    "data": [100, 102, 101, 105, 130, 108, 107],
    "encoder": "json"
  }'
Python SDK
pip install WaveGuardClient

from waveguard_client import WaveGuardClient

client = WaveGuardClient(
    base_url="https://api.emergentphysicslab.com/waveguard"
)

result = client.scan(
    data=[100, 102, 101, 105, 130, 108, 107],
    encoder="json",
)
print(result.is_anomaly)   # True
print(result.score)        # 0.87
print(result.fingerprint)  # Physics fingerprint
Python (requests)
import requests

resp = requests.post(
    "https://api.emergentphysicslab.com/waveguard/v1/scan",
    json={
        "data": [100, 102, 101, 105, 130, 108, 107],
        "encoder": "json",
    },
)
result = resp.json()
print(result["is_anomaly"])  # True
print(result["score"])       # 0.87

Pay Per Scan

AI AGENTS

x402 Micropayments

$0.001 / scan
  • USDC on Base L2
  • Pay-per-request (no subscription)
  • All 37 endpoints
  • Built for AI agent workflows
  • x402 protocol standard

Direct REST API

Free / open access
  • No API key required
  • Rate-limited (fair use)
  • All endpoints available
  • Swagger UI included
  • MCP integration ready

Anomaly detection without the ML pipeline.

No training. No labels. No model drift. Just physics.
Send your first scan now.

Read the Docs