whisper

Speech-to-text transcription converts spoken audio into written text using OpenAI's Whisper model. You simply provide an audio file like 'meeting_recording.mp3' and the Skill outputs the transcribed text in the original language.

Language identification automatically detects the spoken language of an audio recording. The Whisper model supports language identification for 99 languages during the audio processing workflow.

Audio translation converts spoken foreign language audio directly into English text. The Whisper model supports translating audio from 99 languages into English while transcribing the speech.

Whisper model supports various sizes from tiny to large for speech recognition. You select the appropriate model size to balance transcription accuracy and processing efficiency for your audio processing needs.

Faster audio transcription requires Python libraries including openai-whisper, faster-whisper, torch, and transformers. These dependencies enable efficient execution of speech recognition and translation workflows.

Generating subtitles for multilingual video involves transcribing the audio track using the Whisper model. The Skill performs speech-to-text transcription in 99 languages, making audio content accessible and searchable for subtitle generation.