Skip to content

Aalwattar/ParrotInk

Folders and files

NameName
Last commit message
Last commit date

Latest commit

ย 

History

682 Commits
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 

Repository files navigation

ParrotInk Logo

๐Ÿฆœ ParrotInk: Stop Typing. Start Thinking.

License: MIT Latest Release Platform

High-accuracy, ultra-low latency real-time voice-to-text โ€” Native for Windows.

You think faster than you type. ParrotInk closes that gap. It is the only open-source, Windows-native application that brings professional-grade, real-time voice-to-text directly to your cursor. With a polished visual identity and high-fidelity tray feedback, ParrotInk streams your words instantly so you never lose your flow state.


๐Ÿ“– Table of Contents


๐Ÿš€ The Story Behind ParrotInk

I built ParrotInk because existing dictation tools broke my concentration. They either forced me to wait for "batch processing" to dump a giant block of text onto the screen, orโ€”like the built-in Windows dictation (Win + H)โ€”struggled heavily with slight accents.

I wanted true real-time streaming with professional-grade accuracy, and I wanted the freedom to use my own API keys from industry leaders like OpenAI and AssemblyAI. When I couldn't find a single open-source, Windows-native tool that met these standards, I decided to build it. ParrotInk is the result: a tool designed for people who want to type as fast as they can think, without the wait.


โœจ Key Features

  • Zero-Latency Feel: Real-time streaming ensures your text appears as you speak, keeping you in the zone.
  • Global Language Support: Works with almost any language (English, Spanish, French, Chinese, Japanese, etc.) via world-class AI models, so you can work in your native tongue.
  • Native RTL Support: Full support for Arabic and other Right-to-Left languages with correct character shaping and HUD alignment.
  • Smart Injection: Automatically stops recording if you start typing manually, preventing frustrating text collisions.
  • Secure by Design: API keys are stored safely in the Windows Credential Manager, protecting your sensitive credentials.

๐Ÿ–ผ๏ธ Visual Preview

Real-time English & RTL (Arabic) Support

ParrotInk natively handles mixed-language environments and Right-to-Left (RTL) text.

HUD Examples

(Note: Screenshots show the Skia-powered HUD with Acrylic blur and RTL alignment.)


๐Ÿ“ฅ Download & Install

ParrotInk is a portable application. There is no installer needed; just download the EXE and run it.

  1. Download: Click the button below to get the latest version. Download ParrotInk

โš ๏ธ Note on Windows Security: Because ParrotInk is a new open-source project and is not yet signed with a paid Microsoft Developer certificate, you may see a "Windows protected your PC" warning.

To run the app: Click More info -> Run anyway.

As this is 100% open-source software, you can always inspect the code or build the EXE yourself from source if you prefer.

  1. Launch: Open ParrotInk.exe. You will be greeted by a Welcome Onboarding window that explains the basics. You will also see a new icon in your System Tray (the small icons area next to your clock).
  2. Setup API Key:
    • What is an API Key? Think of it as a "secret password" that allows ParrotInk to securely talk to the AI models (OpenAI or AssemblyAI) that handle the transcription.
    • How to get one: Sign up for a free account at AssemblyAI or OpenAI and copy your "API Key" from their dashboard.
    • Where to put it: Right-click the ParrotInk tray icon and select Settings > API Credentials.
  3. Start Dictating: Press Ctrl + Alt + V (default) and start speaking. Your words will appear instantly at your cursor!

Pro Tip: To have ParrotInk start automatically with your computer, right-click the tray icon and check Settings > Run at Startup.


๐Ÿ’Ž Choose Your Engine

ParrotInk requires you to bring your own API key. You only pay the AI providers for the raw milliseconds of audio they process.

๐Ÿ† AssemblyAI (Best for Professional English & Speed)

Featuring Universal-3 Pro (u3-rt-pro) and Instructional Prompting.

  • Why we recommend it: The latest U3 Pro model is the industry benchmark for streaming accuracy. It supports Instructional Prompting, allowing you to guide the AI (e.g., "Always use medical terminology" or "Format as a list").
  • Get Started for $0: New users currently receive $50 in free credit upon signing up.
  • Advanced Features: Supports Shaded Partials in the HUD, letting you see the AI "thinking" in real-time before finalizing the text.
  • Setup: Paste your key from the AssemblyAI Dashboard into the ParrotInk menu.

๐ŸŸข OpenAI (The Global & RTL Powerhouse)

Unmatched support for 50+ languages and Arabic.

  • Why use it: For native Arabic, Japanese, Chinese, or multilingual workflows, OpenAI provides world-class coverage.
  • Language Support: Perfect for users who switch between global languages frequently.
  • Setup: Requires a key from the OpenAI Platform. Note: OpenAI requires a small pre-paid balance (usually $5 minimum) to activate their Realtime API.

๐Ÿ’ฐ What does it cost?

ParrotInk is open-source and 100% free software. Your only cost is your direct usage with the API providers (OpenAI or AssemblyAI), meaning there is no middleman markup.

  • Casual User: A few quick Slack replies a day? Roughly $0.50/month.
  • Power User: 2 hours of dictation every day? Roughly $5.00 - $7.00/month.

๐Ÿ›ก๏ธ Privacy & Security

As an open-source project, transparency is our priority:

  • Encrypted Storage: Your API keys are stored directly in the Windows Credential Manager, encrypted at the OS level.
  • No Local Audio Storage: Audio is streamed via encrypted WebSockets to the provider and is never saved to your hard drive.
  • Local Logs & Stats: All diagnostic data is kept strictly on your machine:
    • Logs: %LOCALAPPDATA%\ParrotInk\Logs\parrotink.log (Rotational, capped at 30MB).
    • Statistics: %APPDATA%\ParrotInk\Stats\stats.json.

๐ŸŽจ Visual Feedback: Understanding the Icon

ParrotInk stays out of your way in the System Tray (the icons next to your clock). The icon color tells you exactly what the app is doing:

  • ๐Ÿ”˜ Grey: Idle โ€” App is ready and waiting for your hotkey.
  • ๐Ÿ”ต Blue: Listening โ€” Capturing your voice and streaming text.
  • ๐ŸŸก Yellow: Connecting โ€” Establishing a secure link to the AI.
  • ๐Ÿ”ด Red: Error โ€” Something is wrong (usually a missing API key).

๐Ÿ” User Manual: How it Works

โšก Simple Use (The "3-Step Flow")

  1. Click into the application where you want to type (Notion, Slack, Word, etc.).
  2. Press Ctrl + Alt + V (or your custom hotkey).
  3. Speak. Your words will appear instantly. Stop speaking or press any key to finish.

๐ŸŽฎ Operation Modes

ParrotInk monitors a global hotkey via native Win32 hooks. You can switch between two distinct modes via the tray menu (Settings > Hold to Talk):

1. Hold to Talk (The "Walkie-Talkie")

Best for short bursts, quick replies, and coding snippets.

  • Action: Press and hold your hotkey.
  • Dictate: Speak your mind.
  • Stop: Release the hotkey. The session ends immediately when you let go.

2. Toggle Mode (The "Hands-Free")

Best for long-form writing, drafting long emails, or deep thinking.

  • Action: Press the hotkey once to start recording. You can now take your hands off the keyboard and focus entirely on your speech.
  • Stop: Press the hotkey again to finish.
  • Smart Stop: If you start typing manually on your keyboard while it's listening, ParrotInk will automatically stop the session.

๐Ÿ› ๏ธ Customization & Advanced Settings

โŒจ๏ธ Changing your Hotkey

  1. Right-click the tray icon.
  2. Select Settings > Change Hotkey....
  3. Press the new key combination (e.g., Alt + S or Ctrl + Space). ParrotInk saves it instantly.

โš™๏ธ Deep Configuration

ParrotInk follows a Portable-First configuration strategy. It looks for its settings in the following order:

  1. Portable Mode: A config.toml file located in the same folder as the ParrotInk.exe. (Ideal for USB drives or custom installs).
  2. Standard Mode: %APPDATA%\ParrotInk\config.toml (Used if no local file is found).

Available Tuning:

  • HUD Styles: Customize the Skia-powered HUD appearance.
  • Acoustic Profiles: Switch between Headset (near-field) and Laptop Mic (far-field) profiles.
  • Latency Tuning: Choose between Fast, Balanced, or Accurate profiles to match your speaking pace.

Detailed documentation on all technical settings can be found in the config.example.toml file.


๐Ÿ—๏ธ Building from Source

ParrotInk is built with Python 3.12+ and uses modern, high-performance libraries like skia-python for rendering and sounddevice for low-latency audio capture.

  1. Prerequisites:
    • Python 3.12+
    • uv (for dependency management)
  2. Clone & Setup:
    git clone https://github.com/Aalwattar/ParrotInk.git
    cd ParrotInk
    uv sync
  3. Run Locally:
    uv run main.py
  4. Build Frozen EXE (Using PyInstaller):
    .\scripts\build_onefile.ps1

๐Ÿค Contributing & Community

ParrotInk is actively maintained and welcomes contributions!

  • Found a bug? Open an issue on the GitHub Issue Tracker.
  • Have an idea? Submit a feature request or open a Pull Request.
  • Code Style: We use ruff for linting and formatting, and mypy for strict type checking. See the local CI scripts for our "Definition of Done".

โš–๏ธ License

Distributed under the MIT License. See LICENSE for more information. Built for speed. Built for Windows. Built for you.

About

Professional-grade, real-time voice-to-text for Windows. Stream your voice directly to any application with ultra-low latency. Supports Arabic (RTL), English, and 50+ languages.

Topics

Resources

License

Stars

Watchers

Forks

Packages

 
 
 

Contributors