Voice Deepfake Detection & Identity Verification API
Classify any voice as real or synthetic, adding the additional security layer your product needs. Integrates via REST API, WebSocket streaming, or native SDK — in less than 3 lines of code.
import requests
response = requests.post(
"https://api.aurigin.ai/v1/predict",
headers={"x-api-key": API_KEY},
files={"audio": open("call.wav", "rb")}
)
print(response.json())
# {"prediction_id": "pred_abc123", "global": {"confidence": 0.99, "result": "spoofed"}, ...}98%+
Detection accuracy
<50ms
Real-time latency
<3
Lines of code to integrate
80+
Languages
98%+
Detection accuracy
<50ms
Real-time latency
<3
Lines of code to integrate
80+
Languages
Endpoints
Two endpoints, complete protection
No ML expertise required. No data stored. Your product gets a trust signal; your users get proof of authenticity and identity.
Deepfake Detection
Classifies any audio input as real or synthetic. Detects voice clones, text-to-speech, and AI-generated speech — in real time or from files. Returns verdict, confidence score, and composite risk score you can threshold to your own tolerance.
Voice Identity Verification
Confirms speaker identity through voice biometrics and cryptographic audio signing. Proves the voice is real AND belongs to the claimed identity. Augments your existing biometric engine — doesn't replace it. Returns match status, confidence, and audit trail reference.
Key capabilities
Built for production
1s
Min. Audio
Minimum speech duration required
Zero
Data Stored
Privacy by design — no audio retained
Robust
All Conditions
Low quality, noise, and all cloning tools
Low
Compute
Runs on a small CPU like a Raspberry Pi
Deployment
Two ways to deploy
Cloud or on-premise. Both production-ready. Both under 3 lines of code.
Cloud API
Secure Aurigin-hosted API. No compute implications for your organization. No installation required. Less than 3 lines of code to integrate.
On-Premise Docker
Off-the-shelf Docker container on your private cloud. No data leaves your servers. Full data sovereignty. Internal API access with <3 lines of code.
Add voice trust to your product
Production-ready deepfake detection and identity verification. Less than 3 lines of code.