Parakeet

Skill tdimino/claude-code-minoan/skills/integration-automation/parakeet

A curated ~/.claude/ configuration for professional development workflows — 90+ skills, 46 hooks, and CLI tools

Install

npx -y skills add tdimino/claude-code-minoan --skill parakeet

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

What its author says it does

Copied from the file, not written here

Local speech-to-text via Handy app (push-to-talk) and NeMo CLI scripts. Parakeet V3: 25 languages, auto-detection, ~30x realtime on M4 Max, 6% WER. This skill should be used when transcribing audio files or dictating voice input.

SKILL.md

5.9 KB, ~1.6k tokens by cl100k_base, as published. Nobody here has run it

Parakeet Dictation Skill

Local speech-to-text powered by NVIDIA Parakeet TDT 0.6B V3 (~600MB model, 100% offline).

Two Modes

1. Handy App (Primary — Push-to-Talk into Any Text Field)

Handy is a free, open-source Tauri app (Rust + React) providing push-to-talk dictation with Parakeet V3 built in. Inference via transcribe-rs (ONNX Runtime, int8 quantized).

brew install --cask handy

Default hotkey: ⌥Space (Option-Space) on macOS, Ctrl-Space on Windows/Linux
Modes: Push-to-talk (hold) or toggle (press to start/stop)
Select Parakeet V3 in Settings → Models (auto-downloads ~478MB)
Grant microphone + accessibility permissions
Includes VAD (Silero), model management UI
Additional models: Whisper (Small/Medium/Turbo/Large), Moonshine, SenseVoice
Models stored at ~/Library/Application Support/com.pais.handy/models/

2. CLI Scripts (Claude Code File Transcription & Terminal Dictation)

CLI scripts remain for headless/terminal use within Claude Code. These use NeMo/PyTorch.

Performance

System	Speed	Engine
Handy (M4 Max)	~30x realtime	transcribe-rs / ONNX int8
Handy (Zen 3)	~20x realtime	transcribe-rs / ONNX int8
Handy (Skylake i5)	~5x realtime	transcribe-rs / ONNX int8
NeMo CLI (MPS)	Varies	NeMo / PyTorch

Accuracy: 6.05% WER (Word Error Rate)
Languages: 25 European languages with automatic detection (no prompting)
Privacy: 100% local processing, no cloud API
License: CC BY 4.0 (model), MIT (Handy app)

Commands

Transcribe Audio File

/parakeet path/to/audio.wav
/parakeet ~/recordings/interview.mp3
/parakeet meeting.m4a

Supported formats: .wav, .mp3, .m4a, .flac, .ogg, .aac

Live Dictation (Terminal)

/parakeet
/parakeet dictate

Record from microphone until Enter is pressed, then transcribe.

Check Installation

/parakeet check

Verify Parakeet is properly installed and model can load.

Setup

Handy (Push-to-Talk UI)

brew install --cask handy

Launch from Applications, select Parakeet V3 model, configure hotkey.

CLI Scripts (Prerequisites)

Parakeet Dictate repo at ~/Programming/parakeet-dictate/ with Python venv

Install dependencies:

cd ~/Programming/parakeet-dictate
uv venv && uv pip install -r requirements.txt

(Optional) Set custom path: export PARAKEET_HOME=/path/to/parakeet-dictate

Implementation

When this skill is invoked:

For audio files: Run the transcription script

cd ~/.claude/skills/parakeet/scripts && \
${PARAKEET_HOME:-~/Programming/parakeet-dictate}/.venv/bin/python transcribe.py "<filepath>"

For live dictation: Run the dictation script

cd ~/.claude/skills/parakeet/scripts && \
${PARAKEET_HOME:-~/Programming/parakeet-dictate}/.venv/bin/python dictate.py

For checking setup: Run the check script

cd ~/.claude/skills/parakeet/scripts && \
${PARAKEET_HOME:-~/Programming/parakeet-dictate}/.venv/bin/python check_setup.py

3. Server Mode (OpenAI-Compatible STT for Takopi)

An OpenAI-compatible transcription server for integration with Takopi and other API clients. Exposes POST /v1/audio/transcriptions wrapping the NeMo CLI.

uv run --with fastapi,uvicorn,python-multipart \
  ~/.claude/skills/parakeet/scripts/parakeet_server.py --port 8384

Takopi config (~/.takopi/takopi.toml):

voice_transcription_base_url = "http://localhost:8384/v1"
voice_transcription_api_key = "local"
voice_transcription_model = "parakeet-tdt-0.6b"

Endpoints: /v1/audio/transcriptions (POST, multipart), /v1/models (GET), /health (GET). Supports response_format: json, text, verbose_json. Accepts .wav, .mp3, .m4a, .flac, .ogg, .aac, .oga, .webm. 50 MB upload limit. 120s transcription timeout.

Model Caches

System	Cache Location	Size	Engine
Handy	`~/Library/Application Support/com.pais.handy/models/`	~478MB	transcribe-rs (ONNX int8)
NeMo CLI	`~/.cache/nemo/`	~1.2GB	NeMo / PyTorch

Model caches are separate. Handy's Parakeet V3 int8 model structure:

parakeet-tdt-0.6b-v3-int8/
├── encoder-model.int8.onnx
├── decoder_joint-model.int8.onnx
├── nemo128.onnx (audio preprocessor)
└── vocab.txt

Troubleshooting

"No module named nemo"

Use the Parakeet virtual environment. Scripts automatically use the correct Python.

"MPS not available"

Apple Silicon Metal acceleration requires PyTorch 2.0+. Falls back to CPU automatically.

"Permission denied: microphone"

Grant microphone access in System Preferences → Privacy & Security → Microphone.

Model download slow

The Parakeet model downloads on first use (~478MB for Handy, ~1.2GB for NeMo). Subsequent runs use cache.

Configuration

Variable	Default	Description
`PARAKEET_HOME`	`~/Programming/parakeet-dictate`	Parakeet Dictate installation path

Dependencies

Handy: brew install --cask handy (standalone, no other deps)

CLI scripts require:

Parakeet Dictate repo at $PARAKEET_HOME (default: ~/Programming/parakeet-dictate)
Python virtual environment at $PARAKEET_HOME/.venv
NeMo toolkit with ASR support (nemo_toolkit[asr]>=2.0.0)
PyTorch 2.0+ (for MPS/CUDA acceleration)
soundfile and sounddevice for audio handling