Personal Project · 2024-07-27
GetDict
AI-based hotkey transcription application for quick, seamless voice-to-text input across any active application window.
Problem
Switching to a dictation app, copying text, and pasting back breaks flow — especially for developers, writers, and operators who need fast capture without leaving their current tool.
Who it serves
Power users who want ambient voice input across any desktop application with minimal friction.
Ownership
Application design, hotkey capture, transcription pipeline, cross-window text insertion, and local/cloud mode configuration.
Architecture
- Global hotkey listener — activates recording without focus change
- Speech capture — audio input with configurable quality
- AI transcription — Whisper-based speech-to-text with custom vocabulary support
- Text insertion — injects transcribed text into the active application window
Technologies
Python · SpeechRecognition · pynput/keyboard · Tkinter/PyQt · OpenAI Whisper
Constraints
- Cross-application text insertion varies by OS and target application behaviour
- Local vs cloud transcription modes trade privacy against accuracy and latency
- Custom vocabulary improves domain-specific accuracy (technical terms, names)
Outcome
A practical productivity tool demonstrating desktop automation, AI integration, and user-centred interaction design — applicable to assistive tooling and operational capture workflows.
Related services
AI applications · workflow automation · desktop tooling