Overview:
Amical is an open-source, local-first AI dictation application. It uses Whisper for on-device speech-to-text and open-source LLMs for intelligent text processing, ensuring complete privacy as everything runs on your machine. A key differentiator is context-aware dictation, where Amical automatically detects the active application—whether you are drafting an email, chatting on Discord, writing code in an IDE, or messaging friends—and formats your speech output accordingly.
Core Features:
Context-Aware Dictation: Automatically adapts speech formatting based on the currently active application, such as email clients, chat apps, or IDEs.
Local-First & Offline: Operates entirely on the user's machine for full privacy, with one-click setup of local AI models included in the application.
Voice Macros & Hotkeys: Supports extensible workflows via custom hotkeys, voice macros, and user-defined automation.
Floating Widget: Includes a minimal, always-on-top widget for frictionless start and stop of dictation sessions.
Use Cases:
Developers: Dictating code comments, prompts, or documentation directly into an IDE with context-aware formatting.
Knowledge Workers: Drafting emails, messages, and documents without switching context, allowing for hands-free note-taking.
Personal Productivity: Creating voice notes and having the content intelligently processed for summaries or tasks.
Why It Matters:
Amical provides a privacy-centric alternative to cloud-based dictation services by processing all audio and text locally using open-source models. Its active app detection differentiates it from generic dictation tools, allowing for a more seamless, context-sensitive workflow without uploading data to external servers. The application is built on a modern tech stack (Electron, Next.js) and includes extensibility options like hotkeys and voice macros for power users.
