Macparakeet
Fast, local-first voice app for Mac. Dictation and transcription powered by Parakeet TDT on the Neural Engine.
Install / Use
/learn @moona3k/MacparakeetREADME
MacParakeet runs NVIDIA's Parakeet TDT on Apple's Neural Engine via FluidAudio CoreML. It has three co-equal modes: system-wide dictation, file/URL transcription, and meeting recording. All speech recognition happens on your Mac.
What it does
Dictation — Press a hotkey in any app, speak, text gets pasted. Hold for push-to-talk, double-tap for persistent recording. Works system-wide.
File transcription — Drag audio or video files, or paste a YouTube URL. Full transcript with word-level timestamps, speaker labels, and export to 7 formats (TXT, Markdown, SRT, VTT, DOCX, PDF, JSON).
Meeting recording — Capture microphone + system audio together, watch a live transcript preview while recording, and save the finished meeting directly into the library with summaries, chat, search, and export support.
Text cleanup — Filler word removal, custom word replacements, text snippets with triggers. Deterministic pipeline, no LLM needed.
AI features — Optional transcript summaries and chat. Ships with a prompt library, built-in and custom summary prompts, multi-summary tabs, and queued summary generation. Use Claude Code or Codex via Local CLI, or connect OpenAI, Anthropic, Google Gemini, Ollama, or OpenRouter. Entirely opt-in, with a fully local setup available when you stay on local providers/features.
Concurrent by design
- Dictation and meeting recording can run at the same time.
- Audio capture stays independent per flow.
- All STT work routes through one shared runtime owner and explicit scheduler with two default slots: a reserved dictation slot and a shared meeting/batch slot, so file transcription can wait behind interactive work.
Performance
- ~155x realtime — 60 min of audio in ~23 seconds
- ~2.5% word error rate (Parakeet TDT 0.6B-v3)
- ~66 MB working memory per active Parakeet inference slot
- 25 European languages with auto-detection
Limitations
- Apple Silicon only (M1/M2/M3/M4)
- Best with English — supports 25 European languages but accuracy varies
- No CJK language support (Korean, Japanese, Chinese, etc.)
Get it
Download: Grab the notarized DMG or visit macparakeet.com. Drag to Applications, done.
First launch downloads the speech model (~6 GB) plus speaker-detection assets (~130 MB) when that default-on feature remains enabled. After that, dictation, meeting recording, and transcription work fully offline.
Build from source:
git clone https://github.com/moona3k/macparakeet.git
cd macparakeet
swift test
scripts/dev/run_app.sh # build, sign, launch
The dev script creates a signed .app bundle so macOS grants mic and accessibility permissions. Set DEVELOPMENT_TEAM=YOUR_TEAM_ID if needed.
CLI:
swift run macparakeet-cli transcribe /path/to/audio.mp3
swift run macparakeet-cli models status
swift run macparakeet-cli history
Tech stack
| Layer | Choice | |-------|--------| | STT | Parakeet TDT 0.6B-v3 via FluidAudio CoreML (Neural Engine) | | STT orchestration | Shared runtime + explicit scheduler across dictation, meeting recording, and transcription | | Language | Swift 6.0 + SwiftUI | | Database | SQLite via GRDB | | Auto-updates | Sparkle 2 | | YouTube | yt-dlp | | Platform | macOS 14.2+, Apple Silicon |
Privacy
All speech recognition runs on the Neural Engine. Your audio never leaves your Mac.
- No cloud STT. The model runs on-device. No audio is transmitted.
- No accounts. No login, no email, no registration.
- Opt-out telemetry. Non-identifying usage analytics and crash reporting go to a self-hosted endpoint only when telemetry is enabled. No persistent IDs, no IP storage, and no transcript/audio content is transmitted. Source code is right here — verify it yourself.
- Temp files cleaned up. Audio deleted after transcription unless you save it.
What does use the network: AI summaries and chat connect to configured LLM providers or CLI tools when you choose them. Sparkle checks for app updates. YouTube transcription downloads video via yt-dlp. Telemetry and crash reports go to our self-hosted server unless you opt out. Core dictation and transcription remain fully offline, and the app supports a 100% local setup when you use only local features/providers.
Note: Builds from source also send telemetry by default. Opt out in Settings or set MACPARAKEET_TELEMETRY_URL to override.
Contributing
- Report bugs — Open an issue
- Submit a PR — Fork, make changes,
swift test, open a PR - Read the specs — Architecture decisions and feature specs live in
spec/
For larger changes, open an issue first.
Support
MacParakeet is free and open source. If it's useful to you, consider sponsoring.
License
GPL-3.0. Free software. Full license.
Made for people who think faster than they type.
Related Skills
node-connect
351.4kDiagnose OpenClaw node connection and pairing failures for Android, iOS, and macOS companion apps
frontend-design
110.7kCreate distinctive, production-grade frontend interfaces with high design quality. Use this skill when the user asks to build web components, pages, or applications. Generates creative, polished code that avoids generic AI aesthetics.
openai-whisper-api
351.4kTranscribe audio via OpenAI Audio Transcriptions API (Whisper).
qqbot-media
351.4kQQBot 富媒体收发能力。使用 <qqmedia> 标签，系统根据文件扩展名自动识别类型（图片/语音/视频/文件）。
