NVIDIA® Riva is a collection of GPU-accelerated multilingual speech and translation microservices for building fully customizable, real-time conversational AI pipelines. Riva includes industry-leading automatic speech recognition (ASR), text-to-speech (TTS), and neural machine translation (NMT) and is deployable in all clouds, data centers, at the edge, and on embedded devices. With Riva, organizations can add speech and translation interfaces that transform chatbots into engaging, expressive multilingual voice AI agents or avatars.

Features

• Multilingual Speech Recognition: Supports automatic speech-to-text recognition and speech-to-text translation, adding punctuation and capitalization, and supporting translation.
• Text-to-Speech (TTS): Converts text into audio with natural-sounding, multilingual speech, customizable with additional, brand-specific voices.
• Neural Machine Translation (NMT): Provides high accuracy in translating text between multiple languages.
• Flexible Deployment: Deploy anywhere—in data centers, on premises, in the cloud, at the edge, or in embedded devices.
• Customizable Pipelines: Customize ASR pipelines for different languages, accents, domains, vocabulary, and context, and TTS pipelines for brand voice and intonation.

Benefits

• High Accuracy: Achieve high multilingual transcription and translation accuracy with state-of-the-art models pretrained on thousands of hours of audio.
• Expressive Voice Generation: Provide out-of-the-box, expressive, professional female and male voices.
• Enterprise-Grade AI: Accelerate the development and deployment of production-grade, multilingual, voice-enabled AI applications.
• Consistent Experiences: Provide consistent experiences to hundreds of thousands of concurrent users with higher inference performance than existing technology.
• Customizable Solutions: Fully customizable across ASR and TTS pipelines for the best possible accuracy and brand voice.