transcribe

Transcribes local audio files to text using NVIDIA Parakeet (NeMo).

Updated Nov 17, 2025

Search Tags:#CLI #audio #local #speech-to-text #transcribe #Parakeet #NeMo

Ready to run

npx skills add https://github.com/ci/persops --skill transcribe-ci

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: transcribe
Source: https://github.com/ci/persops/tree/main/modules/ai/skills/transcribe
Command: npx skills add https://github.com/ci/persops --skill transcribe-ci

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill provides fast, local transcription of audio into text using NVIDIA Parakeet (NeMo, GPU-accelerated) when available, enabling offline speech recognition without cloud services.

Core Features & Use Cases

Local transcription: Convert audio files into accurate text on your machine.
GPU-accelerated: Leverages NVIDIA Parakeet when CUDA is available, with a CPU fallback.
Use Case: Transcribe meetings, podcasts, and lectures for searchability and archival notes.

Quick Start

Transcribe an audio file locally: transcribe /path/to/audio.ogg transcribe /path/to/audio.wav --output /tmp/out.txt

Frequently Asked Questions about transcribe

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

FAQPage Schema

How do I transcribe audio to text locally without using cloud services?▼

You can transcribe audio to text locally by using the NVIDIA Parakeet NeMo model, which runs offline speech recognition directly on your machine. It supports local audio files and outputs text to a specified file.

What is the best way to convert podcast audio files into searchable text?▼

The best way to convert podcast audio is using local speech-to-text with NVIDIA Parakeet, which processes audio files into accurate text for archival notes and searchability. It handles multiple audio formats via ffmpeg.

Do I need an NVIDIA GPU to run local speech-to-text with Parakeet?▼

You do not need an NVIDIA GPU to run local speech-to-text with Parakeet. The system auto-selects CUDA-enabled GPU acceleration when available, but automatically falls back to CPU processing if no GPU is detected.

Can I configure the speech-to-text model used for audio transcription?▼

You can configure the speech-to-text model used for audio transcription by setting the PARAKEET_MODEL variable. This allows you to select different NVIDIA NeMo models for your local transcription tasks.

What audio formats are supported for local transcription?▼

Local transcription supports multiple audio formats processed via ffmpeg, including WAV and OGG files. It handles various input formats for meeting transcription, podcast conversion, and archival audio processing.