Open-source speech recognition software running locally
OpenWhispr is an open-source tool for converting speech into text. It runs entirely locally – simply press and hold the shortcut key to speak, and release it to have the formatted text inserted at the current cursor position. It can be used for writing emails, sending messages, or providing prompts for large language models.
According to the official statements, the speed is 3 to 5 times faster than that of traditional typing methods. Users can choose their preferred models or integrate their own APIs into the system. In addition to dictation functions, this tool can also be used for taking meeting notes – without the need for robots to enter the meeting room. The desktop version is available for macOS, Windows, and Linux, and now an iOS version has also been released. Just like the desktop versions, the iOS version offers real-time transcription capabilities and is completely free to use.
It is ideal for users who value their privacy and prefer not to store their recordings in the cloud. It also serves as a free alternative to paid transcription tools like Wispr Flow. The quality of Chinese language recognition depends on the model you choose; it is recommended that you try different models before making a decision.