Main features¶
das-Peak Streaming is a service to process audio streaming and requests in real-time to register and verify subjects’ identities. It is a cloud service that supports several audio stream protocols in real-time using its API. das-Peak Streaming uses the das-Peak microservice to analyze the audio samples using Veridas proprietary Voice Biometrics AI.
The main features of the das-Peak Streaming service are:
- Real-time Voice biometrics: Designed to verify an identify while user speaks in a natural way, without having to go through a specific identification process.
- Contact Center platform-agnostic: Although Contact Center use case scenario is the most common one, das-Peak Streaming is not limited to Contact Center solutions. It is also possible to send the audio directly without a 3rd Party (no Contact Center required).
- Multiple audio consumers: das-Peak Streaming implements multiple audio consumers based on different audio protocols in order to cover a broad scope:
Genesys Audiohook
: Genesys Cloud websocket-based audio protocol. See Reference.Twilio Media Streams
: Twilio's websocket-based audio protocol. See Reference.- Other audio platforms can be integrated with das-Peak Streaming by sending audio via a subset of Twilio Media Streams websocket specification.
- Extremely easy to integrate: Veridas provides Call Center integral solutions to ease das-Peak Streaming integration process.
- Automated speech extraction: You only have to produce audio in your infrastructure, we have our own consumers to process it.
- Automatic audio quality assessment: das-Peak Streaming has the ability to measure the quality of the audio received and notify format errors and noisy samples.
- Multiple concurrent stream support: With das-Peak Streaming it is possible to use any audio consumer-available. You will be free to use multiple audio protocols, instead of getting tied up to only one.
- Based on das-Peak voice biometrics authentication service: See Reference.
- Ultra-short registration and authentication processes: the technology is capable of generating a biometric voiceprint with only 3 seconds of speech, requiring only 3 seconds of speech for the following authentications.
- Text-independent: The user can speak naturally without the need of repeating a pre-defined sentence.
- Language-independent: The user can be authenticated in any language without prior training of the biometric engine.
- Ultra-fast authentication: The verification is performed in only 140 milliseconds.
- Strict data privacy standards: The cloud service provided does not need to keep or store any user data. GDPR & CCPA compliant.