OpenWhispr

Knowledge workers who type into chat apps, docs, email and code editors are limited by typing speed, often around 40 WPM, which slows down communication and work. People taking meeting notes typically rely on bots or manual note-taking, raising privacy and data-handling concerns, especially for teams in regulated industries subject to HIPAA, GDPR, or SOC 2 requirements. Existing voice dictation tools like Aqua Voice or WisprFlow may require constant connectivity or send data to the cloud, which is a problem for users who travel, lack WiFi, or whose IT teams restrict where company data can go. These users need a fast, accurate, private way to convert speech to text without surrendering control of their data.

OpenWhispr is an open-source dictation app, downloadable for free on macOS, Windows, Linux, and iOS, that lets users write roughly 3x faster by speaking instead of typing, reaching about 150 WPM. It supports local, on-device speech-to-text models as well as optional cloud transcription using a user's own API keys (OpenAI, NVIDIA, Cohere, or OpenWhispr's own Orukeet model), with 0% data retention and no training on user transcriptions by default. The product works across more than 100 languages with auto-detection, auto-learns custom words and jargon from user corrections, supports voice commands like