Amical

Amical is an open-source, local-first AI dictation and note-taking desktop app. It converts spoken language into text and applies context-aware formatting based on the active application the user is currently utilizing, such as an IDE, an email client, or a chat application like Discord. The software is deployed as a desktop application for macOS and Windows, providing a system-wide utility for hands-free typing.

The tool runs entirely on the user's machine to ensure complete data privacy. It utilizes the Whisper model for speech-to-text conversion and integrates with Ollama to run open-source large language models locally for intelligent processing. Users interact with the software through a floating widget and custom hotkeys to manage the start and stop of dictation sessions without leaving their current workspace.

Key features

• Local speech-to-text processing using Whisper
• Context-aware formatting that adapts to the active application
• Offline operation for data privacy
• In-app setup for local AI models
• Floating widget for quick start and stop controls
• Custom hotkeys and voice macros for extensible workflows
• Support for system audio and microphone transcription

Built with Electron, Next.js, and TypeScript, Amical is designed for users who require AI-assisted transcription without sending audio data to cloud servers. The technical stack includes TailwindCSS and Shadcn for the interface and Zod for schema validation. The architecture allows for the future integration of the Model Context Protocol (MCP) to enable voice commands that control other installed applications. It targets developers and professionals who need to generate summaries, tasks, and structured notes from voice input while maintaining a local-only data footprint.

Amical is a privacy-centric productivity tool for local AI dictation and transcription.