ASR

You can transcribe spoken audio into text by sending base64-encoded audio or file-based inputs through the z-ai-web-dev-sdk on the backend. This approach supports batch transcription, real-time streaming, and audio-processing workflows across backend services.

Yes, real-time speech-to-text transcription supports base64-encoded audio inputs for backend integrations. This allows you to process common audio formats directly within your backend services without needing file-based storage.

The best way to test speech-to-text transcription locally is by using the optional CLI usage provided for testing. You can start by transcribing a sample file like 'audio.wav' to verify the SDK-based ASR output quickly.

Yes, the z-ai-web-dev-sdk supports batch transcription from common audio formats. It handles base64-encoded audio and file-based inputs, making it suitable for applications requiring transcripts, captions, or voice-driven features.

You need base64-encoded audio for speech-to-text processing when integrating transcription directly into backend services without relying on file-based inputs. This format allows seamless audio-processing workflows and real-time streaming across web interfaces and mobile apps.