What problem does it solve? Converting spoken audio into accurate, timestamped text for subtitles, transcripts, and video pipelines typically requires either slow local models or complex cloud batch jobs. This Skill provides synchronous cloud transcription of local audio files through Azure AI Speech, without Blob storage or native SDK setup. ## Core Features & Use Cases - Fast Transcription API: Synchronous multipart upload of local audio files returning word-level timestamps, phrase confidence, and detected language. - Speaker Diarization & Language ID: Optional speaker labels with configurable max speakers, plus automatic language identification across candidate locales. - Drop-in Schema Compatibility: Output matches the local faster-whisper transcriber schema, so it feeds directly into subtitle generation and downstream transcript consumers. - Use Case: Transcribe a narration MP3 for a video project, then pass the resulting segments and word timestamps to subtitle generation to produce accurately timed cues. ## Quick Start Set AZURE_SPEECH_KEY and AZURE_SPEECH_REGION, then ask the agent to transcribe your audio file with azure_stt and save the segments to your project's artifacts directory.
