Dictation and Hybrid Input and Editing

Using dictation for text input has become much more viable, especially for multilingual users. For example, I switch constantly between German, English, and Spanish.

A few years ago, reliable programs were scarce and struggled with accents, but modern tools like the open‑source OpenWhispr (https://openwhispr.com/) now connect to powerful inference models such as Voxtral 3 Turbo from <Groq.com>, which runs over 200 times faster than real time and even outpaces stenography.

These systems can not only transcribe speech but also forward the text to an LLM for processing based on a keyword‑triggered prompt (e.g., “hey Whisperer”).

This enables on‑the‑fly translation, email drafting, or other tasks.

Nevertheless, the workflow still benefits from a final manual edit, making the combination of dictation and keyboard editing the most efficient hybrid approach.