Handy is a free, open-source app that converts speech to text offline.
Handy is a free, open-source app that converts voice input into text. Simply speak into the microphone and a locally running AI model will automatically transcribe it, allowing you to input text without using a keyboard.
Handy
Access the above website and click 'download handy for windows' to download the installer for Windows. The installer is in EXE format and is 12.6MB in size. Handy is compatible with Windows, macOS, and Linux (x64).
Launch the downloaded installer and click 'Next'.
Click Next.
When the installation is complete, click 'Next'.
Click “Finish”.
Handy will launch. The default AI model used for transcription is NVIDIA's
The setting screen will be displayed. Specify the input device as 'Microphone'.
Then, press the Ctrl and Spacebar keys simultaneously to activate the microphone input, and simply speak into the application you want to input into. When the microphone input is active, a layer will appear at the bottom of the screen.
Below is a video of what I actually typed using Handy. My pronunciation was quite monotone and halting, but it was converted to text properly.
In addition to Parakeet V3, you can also use OpenAI's Whisper model. You can change the model by clicking the model name in the lower left. However, you will need to download the Whisper model. In this example, select the Whisper Large model and click 'Download.'
Once the download was complete, Whisper Large was activated.
Now the listening language is Japanese. You can select 'Japanese' in the Language field.
Below is a Japanese sentence being read out loud and transcribed.
Select 'History' on the left to view the audio that has been recognized and the text that has been transcribed so far.
Developer CJ Pais said he developed Handy because he broke his finger, put a cast on it, and couldn't use his hand. He added that there was no open-source, extensible speech-to-text app available, so he decided to develop it.
Related Posts:
