How do I resolve 'Connection Refused' errors?

This usually means the whisper-server isn't running. You can check the status using 'systemctl status whisper-server' or restart the service as guided by the skill's troubleshooting section.

What is the default model used by Whisper Transcription?

The skill is configured to use the high-accuracy large-v3 model, which provides excellent results for both English and multilingual speech processing.

What are the hardware requirements for this skill?

It is optimized for NVIDIA GPUs with CUDA support. The large-v3 model requires approximately 6GB of VRAM to operate efficiently.

Does this skill require an internet connection?

No, it operates entirely on your local machine using a self-hosted whisper.cpp server, ensuring your audio data remains private and secure.

Can I record and transcribe audio in one workflow?

Yes, the skill provides implementation patterns for recording audio via shell commands like 'arecord' and immediately sending the output to the transcription server.

Whisper Transcription

Name: Whisper Transcription
Author: lawless-m

bylawless-m

データサイエンスとML

Transcribes audio files into text using a local whisper.cpp server with GPU acceleration.

概要

This skill integrates local speech-to-text capabilities into the Claude environment by interfacing with a whisper.cpp server. It allows users to convert audio recordings, voice notes, and media files into high-accuracy text using the large-v3 model. Optimized for NVIDIA GPUs with CUDA support, it enables real-time transcription, shell-based recording, and Python integration, making it an essential tool for developers and researchers who require private, high-performance transcription without relying on external cloud APIs.

主な機能

GPU/CUDA acceleration utilizing the large-v3 model
Built-in troubleshooting for VRAM management and server connectivity
Support for multiple audio formats including WAV and MP3
Local speech-to-text transcription via whisper.cpp
Seamless integration with shell commands and Python scripts
0 GitHub stars

ユースケース

Transcribing recorded voice notes or meeting audio directly from the terminal
Building automated speech-to-text pipelines in Python applications
Generating text transcripts for local media files to assist in documentation

概要

主な機能

GPU/CUDA acceleration utilizing the large-v3 model
Built-in troubleshooting for VRAM management and server connectivity
Support for multiple audio formats including WAV and MP3
Local speech-to-text transcription via whisper.cpp
Seamless integration with shell commands and Python scripts
0 GitHub stars

ユースケース

Transcribing recorded voice notes or meeting audio directly from the terminal
Building automated speech-to-text pipelines in Python applications
Generating text transcripts for local media files to assist in documentation