Since August 2, 2026, Article 50 of the EU AI Act requires AI-generated audio to be marked as such and deepfakes to be disclosed. ORAVYS checks the recordings you receive and keeps a documented trail of each check.
Key dates for AI Act compliance obligations
How ORAVYS addresses each regulatory obligation
| Requirement | Description | ORAVYS |
|---|---|---|
| Detection | Identify AI-generated or manipulated audio content | Three result bands: strong synthetic signals, no strong synthetic signal, or inconclusive |
| Labeling | Mark synthetic content as AI-generated in a machine-readable format | Structured JSON reports with confidence scores |
| Documentation | Maintain records of detection and labeling activities | Audit trail, tier access logs, file fingerprint (SHA-256) and timestamp on each report |
| Robustness | Detection must work across compression, noise, and codec variations | Three result bands: a recording too degraded to decide is reported inconclusive, never forced into a verdict |
| Explainability | Provide meaningful explanations for detection decisions | Each report states its result band and its limits; inconclusive is a valid outcome |
| Scalability | Handle enterprise-volume content moderation | API-first, batch processing, webhook callbacks |
Prices in USD, from the same grid as the pricing page.
100 scans, 1,000 API calls a month. Deepfake detection API. Same result as a single scan, in your own software. Email support. Ideal for content agencies, small BPOs, and podcast platforms.
$99 /month
See API pricing2,000 scans a month, API and webhooks. Art. 50 transparency webhooks. PDF reports. Priority support. For fintechs, call centers, and content platforms.
$899 /month
Book a DemoVolume by contract, for banks, insurers, and governments. Dedicated instance or VPC deployment. Multiple independent checks. Documented, auditable forensic reports. SLA by contract. SIEM integration. White-label option.
from $2,900 /month
Contact SalesOrganizations affected by Article 50 transparency obligations
Screen recorded calls for synthetic caller voices. Protect agents and customers from voice cloning fraud.
Voice authentication verification for banking, insurance claims, and KYC compliance.
Documented, auditable analysis of audio evidence. Methodology oriented for use by qualified legal professionals; admissibility determined per jurisdiction.
Content moderation at scale. Label AI-generated audio before publication or distribution.
Verify candidate voice authenticity in remote interviews. Protect internal communications.
Ensure patient identity verification. Protect telehealth consultations from impersonation.
Get an audit trail and transparency webhooks to support your Art. 50 work. This is not legal advice.
Book a Compliance Demo Request API access