FAQ
Simply enter the text you desire to generate audio from in the text box. Select the speaker on the left of the text box that matches the language your text is from. Different speakers will generate audio with different tones and speech patterns.
Click Generate to generate your audio! Depending on the length of text you may need to wait longer for the generation to complete.
Text and Audio Diversity sliders allow you to finetune the characteristcs of the generated audio. If your audio seems to have unwanted artifacts, turning down Sensitivity may help in alleviating this situation.
Increasing Diversity would make the audio sound less mechanical and closer to natural human speech
If unsure, a good balanced diversity value would be 70%, to give you natural sounding audio with fewer artifacts.
By enabling Preview Mode, Blurt will use a smaller AI Model to give you a way to quickly get a sample to listen to. While the quality may be worse than the larger model, you are able get an idea of how your audio will sound, saving you time if you need to make corrections to your text, adjust diversity sliders or change speakers.