Whisper Audio Transcriber

About

The Audio Transcriber skill integrates WhisperX into your workflow, allowing you to convert speech from various media formats into highly accurate text directly through Claude Code. It features advanced word-level alignment, automatic language detection, and supports multiple output formats including SRT and VTT for subtitles or JSON for structured data. With configurable model sizes ranging from 'tiny' to 'large-v2', users can easily balance processing speed with transcription accuracy depending on their hardware and quality requirements.

Key Features

• High-accuracy transcription using WhisperX machine learning models
• Support for multiple output formats including SRT, VTT, TXT, and JSON
• 0 GitHub stars
• Configurable VAD (Voice Activity Detection) to prevent segment skipping
• Automatic language detection and multi-language support (zh, en, ja, etc.)
• Word-level timestamp alignment for precise subtitle and data synchronization

Use Cases

• Extracting structured audio data with per-word timestamps for AI analysis
• Generating professional SRT or VTT subtitles for video content
• Transcribing meeting recordings or interviews into searchable text documents