ASR

To transcribe speech audio into text in Node.js, run the provided sample script at skills/ASR/scripts/asr.ts with a valid audio file path. It uses the z-ai-web-dev-sdk on a secure backend to process the audio and output the resulting text transcription.

Yes, you can use base64-encoded audio for speech-to-text transcription. This enables the automated transcription of audio and video notes directly through the z-ai-web-dev-sdk without needing to reference external file paths.

Yes, this speech-to-text approach supports both batch and streaming capabilities. This allows you to handle large-scale transcription tasks or process real-time voice input transcripts across web and mobile applications.

Yes, you must install the z-ai-web-dev-sdk in your Node.js environment to run the ASR script. The script relies on this dependency to execute the backend transcription logic and convert spoken audio into text.

The best way to build voice-enabled interfaces is to use the provided back-end integration examples with the z-ai-web-dev-sdk. This approach streamlines transcription tasks and enables voice input features, transcripts, and searchable text across your applications.

The transcription is limited to environments where the z-ai-web-dev-sdk is installed, specifically requiring a Node.js environment. You must also provide a valid audio file path or base64-encoded audio to successfully obtain a transcription.