moona3k/macparakeet

Fast, private, local-first voice app for Apple Silicon Macs — dictation, file/media transcription, meeting recording, Transforms, and a public automation CLI. Free and open-source.

What it solves

MacParakeet is a local-first voice application for Apple Silicon Macs that provides a private, fast way to handle speech-to-text tasks. It eliminates the need to send audio to the cloud for transcription, offering a unified tool for system-wide dictation, recording and transcribing meetings, and converting audio/video files or media URLs into text.

How it works

The app uses on-device speech recognition engines (such as Parakeet via FluidAudio CoreML, WhisperKit, Nemotron, or Cohere Transcribe) to process audio locally on the Mac's Neural Engine. It integrates with macOS system features like EventKit for calendar reminders and yt-dlp for media URL processing. For users who want additional intelligence, it offers opt-in AI features (summarization, chat, and text transforms) that can connect to either local LLM runtimes like Ollama or cloud providers via API.

Who it’s for

Users of Apple Silicon Macs who require high-privacy transcription and dictation, professionals who need to record and summarize meetings, and developers who want a scriptable CLI for voice automation.

Highlights

  • System-wide Dictation: Push-to-talk or hands-free dictation that pastes text directly into any active application.
  • Local-first Transcription: Supports dragging and dropping files, folders, or pasting URLs from platforms like YouTube, X, and TikTok.
  • Podcast Integration: Searches and transcribes Apple Podcasts via the iTunes lookup API.
  • Meeting Recording: Captures system audio and microphone simultaneously with live local transcript previews.
  • Calendar Sync: Reads local macOS Calendar events to provide reminders or automatically start recordings.
  • Automation CLI: A Swift-native command-line interface for managing models, transcribing files, and accessing transcription history.
  • Deterministic Text Cleanup: A non-AI pipeline for removing filler words and expanding custom snippets.
  • Flexible AI Integration: Opt-in support for cloud LLMs or local runtimes (Ollama, LM Studio) for summarizing and chatting with transcripts.

Related

  • Project
    Muesli-HQ/muesli

    A local-by-default dictation and meeting transcription app for macOS that uses on-device AI models for privacy and low-latency speech-to-text.

  • Project
    moinulmoin/voicetypr

    An offline-first voice-to-text dictation app for macOS and Windows that transcribes speech and inserts it directly into any active application.

  • Project
    watzon/pindrop

    An open-source, Mac-native AI dictation app that converts speech to text using local Apple Silicon engines or optional cloud services for maximum privacy.

  • Project
    Gremble-io/Detto

    A local-first macOS app for Apple Silicon that provides private, on-device transcription for meetings, voice memos, and system-wide dictation, outputting structured Markdown files.

  • Project
    FrigadeHQ/yap

    A lightweight, on-device voice dictation tool for macOS (Apple Silicon) that uses native Apple APIs to transcribe speech to text locally and paste it into any active application.