Offline Speech-to-Text: Why I Built ditado Instead of a Subscription

I talk to AI agents all day. Typing responses breaks flow. Existing dictation tools either require a subscription, send your audio to the cloud, or both. So I built ditado — an offline Windows dictation app that never sends your data anywhere.

How It Works

Press F8 to start recording, press F8 again to stop. The transcription runs locally via faster-whisper, and the result is auto-pasted into whatever application you’re using — Notion, Obsidian, ChatGPT, you name it.

Technical Highlights

• Fully offline — faster-whisper runs on CPU with surprisingly good performance on modern hardware
• 13 languages — automatic language detection or manual selection
• System tray app — sits in the background, hotkey works globally
• Zero configuration — install and use, no API keys or accounts

The .exe installer is available. It’s used daily for note-taking and AI interaction — and it costs nothing to run.

Built with: Python, faster-whisper, PyInstaller, sounddevice