Build the Next-Gen
AI Video Models
Integrated model for voice cloning, recognition, synthesis and real-time translation, powering video creation, live streaming, content localization and intelligent voice applications.
Core Model Matrix
Modular and composable voice and audio-video models for scalable content production.
Voice Cloning Model
Clone voice tone, style, and expressive details from reference audio for dubbing, character voices, and branded voice assets.
Speech Generation Model
Generate natural speech from text for voiceovers, narration, courses, audio content, and batch production.
Speech Recognition Model
Robustly recognize names, hot words, conversational speech, and complex audio from livestreams, videos, courses, and interviews.
Real-time Speech Translation Model
Recognize and translate speech in real time for live subtitles, multilingual meetings, video understanding, and live interaction.
Intelligent Subtitle Model
Generate more stable subtitles with speech recognition and context-aware translation for videos, courses, and media localization.
Paralinguistic Understanding Model
Understand not only what is said, but also tone, pauses, speaking rhythm, and conversational expression.
Core AI Capabilities
Built on voice understanding and generation, supporting creation, translation, interaction, and localization workflows.
High-naturalness voice cloning
Accurately clone voice tone, speaking style, and expressive details, making generated speech natural and emotionally rich.
partial editing & efficient refinement
Regenerate selected lines or segments to fix pronunciation, tone, or content without recreating the entire audio.
ultra-fast speech generation
Generate high-quality speech quickly with fast inference, suitable for real-time interaction and large-scale production.
complex speech recognition
Optimized for names, game terms, trending words, conversational speech, and vertical vocabulary in complex scenarios.
streaming end-to-end speech translation
Continuously output recognition and translation results during speech input for live subtitles, meetings, videos, and real-time interaction.
paralinguistic information understanding
Understand tone, pauses, speaking speed, and expression style to make recognition, translation, and generation more natural.
Developer Tools
Access voice cloning, speech recognition, real-time transcription, and streaming translation through APIs and SDKs.
API Access
Integrate voice cloning, speech recognition, speech generation, and real-time translation through standard APIs.
Official SDKs
Use official SDKs for major languages to reduce integration effort and speed up development.
Interactive Playground
Test models online and quickly validate cloning, recognition, translation, and generation results.
Full Platform Ecosystem
Available across mobile, desktop, and browser extensions for creation, translation, and real-time interaction.
