ASR

You can transcribe audio files into text on the backend by reading the file, converting it to base64, and calling the ASR function through the z-ai-web-dev-sdk. This handles the speech-to-text transcription and returns the text payload.

Yes, this ASR Skill supports real-time transcription for voice interfaces. It handles both batch processing of audio files and live processing for streaming applications on the backend using the z-ai-web-dev-sdk.

Before processing audio, you need a backend runtime and the z-ai-web-dev-sdk installed. You must also read your audio files and convert the audio data into base64 encoding to pass into the transcription function.

Yes, this speech-to-text Skill works for batch processing meeting recordings. It transcribes spoken language from audio files into searchable text transcripts, making it suitable for converting recorded meetings into written text.

Base64 encoding is required for audio processing because the z-ai-web-dev-sdk ASR function expects the file data in a base64 string format. This allows the backend to securely transmit the audio payload to the transcription service.