skills.rest
tamagusko
tamaguskoCOMMUNITY
Explore all skills by this author →
WebsiteGitHub

whisper

Transcribe and translate audio in 99 languages using OpenAI's Whisper speech recognition models.

Updated Oct 7, 2022
Search Tags:#translation#speech recognition#transcription#whisper#openai#audio processing#asr
One-click install
npx skills add https://github.com/tamagusko/linux-cfg --skill whisper-tamagusko
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: whisper
Source: https://github.com/tamagusko/linux-cfg/tree/main/dotfiles/claude/skills/whisper
Command: npx skills add https://github.com/tamagusko/linux-cfg --skill whisper-tamagusko

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires openai-whisper, transformers, torch, and includes references (resource) components.

What problem does it solve? Converting audio recordings into text manually is slow and error-prone, especially across multiple languages. This Skill provides guidance for using OpenAI's Whisper model to automate speech-to-text transcription, translation to English, and language identification. ## Core Features & Use Cases - Multilingual Transcription: Transcribe audio in 99 languages with six model sizes (tiny to large) trading speed for accuracy. - Translation & Timestamps: Translate any supported language to English and generate word-level timestamps or SRT/VTT subtitle files. - GPU Acceleration & Batching: Run on CUDA GPUs for 10-20x speedups, batch-process multiple files, or use faster-whisper for streaming. - Use Case: Transcribe a folder of podcast episodes by loading the turbo model, iterating over each MP3 file, and writing the resulting text segments to individual transcript files. ## Quick Start Ask the assistant to transcribe an audio file such as meeting.mp3 using the Whisper turbo model and save the output as a text file with timestamps.

Frequently Asked Questions about whisper

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

FAQPage Schema
How do I transcribe audio to text with Whisper in Python?▼

Install openai-whisper and ffmpeg, then load a model with whisper.load_model("base") and call model.transcribe("audio.mp3"). The result dictionary contains the full text plus timestamped segments you can iterate over.

Which Whisper model size should I use?▼

Use turbo for the best speed-to-quality balance, base for quick prototyping, and large for maximum accuracy on multilingual or low-resource languages. Tiny and base models need about 1 GB of VRAM, while large requires around 10 GB.

Can Whisper translate audio into English?▼

Yes, pass task="translate" to model.transcribe() to convert speech in any of the 99 supported languages into English text. Translation only outputs English; it cannot translate between other language pairs.

Does Whisper support GPU acceleration?▼

Yes, Whisper automatically uses a CUDA GPU when available, giving 10-20x faster transcription than CPU. You can force a device with whisper.load_model("turbo", device="cuda") or device="cpu".

What are the limitations of Whisper transcription?▼

Whisper can hallucinate or repeat text, accuracy degrades on audio longer than 30 minutes, and it provides no speaker diarization. It is also not suitable for real-time live captioning due to latency.

Related Skills

Explore More Software Engineering Skills→
buda-aibuda-ai
25

audio-extractor

Extract and enhance audio from video files for MP3/WAV/AAC output.

Official
Intermediate
davila7davila7
30.5k

rag-engineer

Designs retrieval-augmented generation pipelines with embeddings, chunking, and hybrid search.

Community
Intermediate
lidge-junlidge-jun
5

transcribe

Transcribe audio files into text with speaker diarization using OpenAI's Whisper models.

Community
Intermediate
framerslabframerslab

streaming-stt-deepgram

Transcribe live speech via Deepgram's WebSocket API with speaker diarization.

Official
Advanced