Frequently asked questions
Bhasa.io is a fully secure service. All your data sent between your device and the service is end to end encrypted. The service is completely automated with no manual intervention anywhere. We ensure that no human can access or transfer your data onto other systems.
You can permanently remove the imported audio / video files along with their transcripts from our Service. Once you delete the files you uploaded, please note that there’s no recovering them back.
The generic models whether used by us or by any other service suffer on quality when run on long tail accents. Merely trying to address them with domain specific model does not solve the problem. This is why we encourage you to create Custom models for the audios by providing annotated data.
You can easily have custom models generated for your voice. Broadly, you will be asked to speak out loud the readouts shown to you. This voice will be captured by the page and further used as annotation data to train the custom model.
No, we don’t support that yet.