Free voice dictation that works in any app. Open source, private, and built by a student who needed a tool that actually existed.
Rota AI started the way most useful software does: as a personal need. After trying Wispr Flow and loving it, the 14-day free trial ended, and the $15/month subscription was not something a student could justify.
So the next best option was to build it. Months of reverse engineering, reading research papers on Whisper and voice activity detection, and late nights debugging Windows text injection later, Rota AI was born.
Today, Rota AI is used by developers, writers, students, and accessibility users around the world. It is free, open source, and always will be.
Rota AI is built with Python and PySide6 on the desktop, powered by OpenAI Whisper for transcription (via Groq, Gemini, or local Ollama), Silero VAD for voice activity detection, and an LLM-based cleanup pass for natural, polished output.
The architecture is a 7-stage pipeline running on dedicated threads (audio capture, voice activity detection, transcription, AI cleanup, context detection, text injection, and persistence), ensuring the UI never freezes during processing.
API keys are encrypted at rest using platform-native encryption (DPAPI on Windows, Keychain on macOS). Session history is stored in a local SQLite database that you can clear at any time.
Every line of Rota AI is available on GitHub under the MIT license. You can audit the code, contribute improvements, fork it for your own projects, or just verify that it does exactly what it claims to do and nothing more.
The project is built in the open. Issues, discussions, and pull requests are welcome. No corporate walled garden. No hidden agenda. Just a tool that should exist.
Have questions, suggestions, or want to contribute? Reach out via GitHub or email tl24btcs@gmail.com.