ASR

Converts speech to text using the z-ai-web-dev-sdk with support for base64 encoded audio files. You can either use the CLI for quick transcriptions (z-ai asr --file audio.wav) or integrate the SDK into backend code for production apps. Handles common formats like WAV and MP3, includes caching examples, and ships with batch processing utilities. The documentation is thorough with real code samples for everything from simple file transcription to building a full ASR service class. Backend only, no client-side usage. If you need to transcribe audio files or build voice input features, this covers the basics and gives you working patterns to extend.