Integrate enterprise-grade deepfake detection into your application in less than five minutes. Our obsidian-grade security ensures every frame is analyzed with mathematical precision.
To begin, you'll need a unique authentication token. Head to your Dashboard to generate a production key. Keep this key confidential; it provides root access to your analysis quota.
Use our REST endpoint to submit a video or image for analysis. We recommend using a secure URL for the media_url to ensure integrity during transport.
curl -X POST "https://api.deepfake-obsidian.com/v1/analyze" \ -H "Authorization: Bearer $OBSIDIAN_API_KEY" \ -H "Content-Type: application/json" \ -d '{ "media_url": "https://storage.provider.com/video_sample.mp4", "detection_depth": "neural_surgical", "webhook_url": "https://yourdomain.com/webhooks" }'
The API returns a highly detailed JSON object. Look for the confidence_score and the neural_artifacts array to understand the AI's findings.
{ "job_id": "job_01HF...99Z", "status": "completed", "analysis": { "is_deepfake": true, "confidence_score": 0.9842, "model_version": "obsidian-prime-v4", "neural_artifacts": [ { "type": "lip_sync_mismatch", "severity": 0.92 }, { "type": "frequency_anomaly", "severity": 0.88 } ] } }
Pro Tip: All confidence scores above 0.95 are considered high-confidence anomalies. We recommend manual review for scores between 0.70 and 0.90.
Dive deep into our full parameter documentation and advanced settings.
Explore Reference arrow_forwardLearn how to secure your callbacks and sign requests for peak integrity.
Read Security Docs arrow_forward