Repolode
Submit
amicalhq

amicalhq/amical

Open-source, local-first AI dictation app that turns speech into text on macOS and Windows, working offline with Whisper.

Open-source alternative to:Dragon NaturallySpeakingApple Dictation

View on GitHubWebsite
Stars
1,561
Forks
142
Open issues
62
Last commit
7 days ago

What is it?

Amical is a desktop dictation app that converts speech to text and runs entirely on the user's machine. It uses Whisper for speech-to-text and open source LLMs for post-processing, and it works offline once a local model is downloaded in the app.

The app is context-aware: it detects the active application and formats the dictated text to suit it, whether that is an email draft, a chat message, or a prompt written in an IDE. Global hotkeys and a floating widget let you start and stop dictation without touching the keyboard.

The project is built with Electron, TypeScript, and Turborepo. It currently supports macOS and Windows, with Linux not yet supported, and the README lists voice notes summarization as in progress and meeting transcription and MCP voice commands as planned.

Key features

  • 01Speech-to-text dictation powered by Whisper with optional AI-enhanced accuracy.
  • 02Context-aware formatting based on the active application.
  • 03Works offline with local models that can be set up in-app with one click.
  • 04Global hotkeys and a floating widget for starting and stopping dictation.
  • 05Available for macOS and Windows, installable via Homebrew on macOS.

Use cases

  • Drafting emails or messages by voice when typing is slow or inconvenient.
  • Writing prompts or code comments in an IDE hands-free.
  • Dictating text in sensitive settings where audio and text must stay on the local machine.

Quick start

  1. 1

    Install the app with Homebrew (macOS)

    brew install --cask amical
  2. 2

    Clone the repository with submodules

    git clone --recurse-submodules https://github.com/amicalhq/amical.git
  3. 3

    Install dependencies and download the Node.js binary

    pnpm install --frozen-lockfile
  4. 4

    Start the desktop app in development mode

    pnpm turbo run dev --filter=@amical/desktop

Things to know

  • Desktop app runs only on macOS and Windows; Linux is not supported.
  • Local Whisper transcription requires macOS 15 or later.
  • Building from source needs Xcode or Visual Studio Build Tools, CMake, and a multi-minute native compile.

Topics

#speech-to-text#dictation#whisper#local-ai#offline#electron#transcription#ai#ai-note-taking-app#asr

FAQ

What is amical?
Open-source, local-first AI dictation app that turns speech into text on macOS and Windows, working offline with Whisper. Amical is a desktop dictation app that converts speech to text and runs entirely on the user's machine. It uses Whisper for speech-to-text and open source LLMs for post-processing, and it works offline once a local model is downloaded in the app.
Is amical free to use? What license does it have?
amical is open source under the MIT license. Check the license terms before using it in commercial products.
Is amical still maintained?
Actively maintained. The last commit to amical was on Oct 13, 2026.
What can I use amical for?
Drafting emails or messages by voice when typing is slow or inconvenient. Writing prompts or code comments in an IDE hands-free. Dictating text in sensitive settings where audio and text must stay on the local machine.
How do I install amical?
Install the app with Homebrew (macOS): brew install --cask amical Clone the repository with submodules: git clone --recurse-submodules https://github.com/amicalhq/amical.git Install dependencies and download the Node.js binary: pnpm install --frozen-lockfile Start the desktop app in development mode: pnpm turbo run dev --filter=@amical/desktop
Where can I get amical?
The source code and installation instructions are on GitHub: https://github.com/amicalhq/amical

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects