parakeet-mlx

Transcribe audio to TXT, VTT, or SRT locally with Parakeet MLX.

1|Updated Feb 15, 2026

Search Tags:#transcription #apple silicon #local #speech-to-text #asr #parakeet mlx

Ready to run

npx skills add https://github.com/rustykuntz/rocie-tools --skill parakeet-mlx

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: parakeet-mlx
Source: https://github.com/rustykuntz/rocie-tools/tree/main/skills/parakeet-mlx
Command: npx skills add https://github.com/rustykuntz/rocie-tools --skill parakeet-mlx

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill provides a local, on-device solution for converting spoken audio into text, eliminating the need for cloud-based APIs and associated costs or privacy concerns, specifically optimized for Apple Silicon hardware.

Core Features & Use Cases

Local ASR: Performs speech-to-text transcription directly on your machine using the Parakeet MLX model.
No API Key Required: Operates entirely offline, making it suitable for sensitive data or environments without internet access.
Use Case: Transcribe meeting recordings or voice notes directly on your MacBook Pro without uploading any audio data to external servers.

Quick Start

Use parakeet-mlx to transcribe the audio file located at /path/to/your/audio.mp3 and output the result as plain text.

Frequently Asked Questions about parakeet-mlx

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

FAQPage Schema

How do I transcribe audio locally on Apple Silicon without an API key?▼

Local speech-to-text transcription on Apple Silicon can be done entirely offline using the Parakeet MLX model, converting spoken audio into text without cloud-based APIs or internet access.

What audio and output formats are supported for local speech-to-text transcription?▼

Local speech-to-text transcription supports various audio input formats processed via ffmpeg, and outputs plain text alongside TXT, VTT, and SRT subtitle formats.

Do I need ffmpeg to run local speech-to-text on my MacBook Pro?▼

Yes, ffmpeg is required for audio processing when performing local speech-to-text transcription, and Parakeet MLX models are automatically downloaded from Hugging Face to your cache directory.

Can I transcribe sensitive meeting recordings without uploading audio data to external servers?▼

Yes, on-device speech-to-text transcription processes sensitive audio files entirely offline, eliminating privacy concerns associated with uploading meeting recordings or voice notes to external servers.

What's the best way to generate SRT subtitles from audio files on Apple Silicon?▼

Generating SRT subtitles from audio files on Apple Silicon is best handled locally by the Parakeet MLX model, which processes audio inputs directly on your machine and outputs SRT format.

Why does local ASR transcription require downloading models from Hugging Face?▼

Local ASR transcription downloads Parakeet MLX models from Hugging Face to your cache directory because the speech-to-text processing runs entirely on-device, requiring the model weights to be stored locally for offline execution.