About dIKta.me

What is dIKta.me?

dIKta.me is a voice dictation engine for Windows that runs entirely on your machine. Press a hotkey, speak, and your words appear in whatever app you're using — Word, Slack, your browser, a terminal, anything.

It's not just transcription. A local AI model cleans up your speech, removes filler words, formats your text, and can even follow instructions like “write this as a professional email” or “translate this to Spanish”.

Why does this exist?

Every voice tool out there wants to send your audio to the cloud. They want a monthly subscription. They cap your words. They own your data.

dIKta.me was born from a simple frustration: why should I need an internet connection and a recurring payment just to talk to my computer? Your voice is personal. Your ideas are yours. The processing should happen on your hardware.

How it's built

dIKta.me V2 is a ground-up rewrite in C# with WinUI 3, designed for performance and a native Windows experience.

  • Speech-to-Text: Whisper V3 Turbo running locally via faster-whisper, or cloud streaming via Deepgram
  • AI Brain: Any Ollama-compatible model (Gemma, Llama, Mistral) for formatting, cleanup, and intelligence
  • Text-to-Speech: Kokoro ONNX for fully local voice synthesis
  • Cloud option: Optional wallet credits for Deepgram + Gemini Flash when you want cloud speed without managing API keys

The Power Version is source-available — you can read every line of code and know exactly what's running on your machine.

The developer

dIKta.me is a solo indie project, co-authored with AI tools. No VC funding, no corporate agenda, no growth-at-all-costs mindset. Just one developer who wanted better voice tools and decided to build them.

Support the project

dIKta.me is independent and self-funded. If you want to help keep it that way:

Join the WaitlistSupport on Ko-Fi