High-quality speech-to-text transcription using Zhipu AutoGLM ASR, supporting long audio chunking, context-aware modes, and timestamp segmentation.