whisper

Transcribe audio into text across 99 languages with openai-whisper.

2|2|Updated Apr 16, 2026
One-click install
npx skills add https://github.com/huidge/hermes-skills --skill whisper-huidge
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: whisper
Source: https://github.com/huidge/hermes-skills/tree/main/mlops/models/whisper
Command: npx skills add https://github.com/huidge/hermes-skills --skill whisper-huidge

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

Transcribe audio into text across 99 languages.

Core Features & Use Cases

  • 99 languages supported for transcription and language identification
  • Translation to English and multilingual text processing
  • Multiple model sizes to balance speed and accuracy for different workloads
  • Use cases include podcast transcription, meeting notes, and media archiving

Quick Start

Install the openai-whisper package, load a model, and run transcription on your audio file.

Frequently Asked Questions about whisper

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I transcribe audio files into text across multiple languages?

Audio transcription into text across 99 languages is done by loading a model and running it on your audio file. It supports podcast transcription, meeting notes, and media archiving workflows.

Do I need to install openai-whisper and torch to run speech-to-text transcription?

Yes, openai-whisper, transformers, and torch are required dependencies to run speech-to-text transcription. These libraries provide the underlying model architecture and inference capabilities needed to process audio files.

Can I translate audio from other languages into English text?

Yes, translating audio into English text is supported alongside multilingual text processing. The skill identifies the spoken language from 99 supported languages and can translate the transcription directly into English.

What is the best way to balance speed and accuracy for large audio processing workloads?

Balancing speed and accuracy for audio processing workloads is achieved by selecting from multiple model sizes. Smaller models run faster while larger models provide higher accuracy for complex transcription tasks.

Does multilingual speech recognition work for podcast transcription and meeting notes?

Multilingual speech recognition works effectively for podcast transcription, meeting notes, and media archiving. It processes audio across 99 languages and identifies the spoken language automatically.

What are the limitations when using different model sizes for speech recognition?

Limitations of different model sizes for speech recognition involve a trade-off between processing speed and transcription accuracy. Smaller models consume less memory and run faster but may sacrifice precision on complex or noisy audio files.