Microsoft’s Copilot Can Now Read Aloud Your Scripts
Microsoft has added a new AI feature to Copilot.
It can now generate audio natively.
The feature lets users hand a script to Copilot.
The AI then converts it into a voiceover.
Different voice styles are available.
Mustafa Suleyman, CEO of Microsoft AI, made the announcement.
He shared the news on X (formerly Twitter).
The feature is powered by the MAI-Voice-1 model.
This model was released in August.
Currently, the feature is available only in Copilot Labs.
A personal account sign-in is required.
There are three audio modes.
The first is Scripted.
It reads text word-for-word.
It is best for formal use.
The second is Emotive.
It sounds dramatic and flashy.
It is ideal for marketing and ads.
The third is Story.
It uses multiple voices and styles.
It is useful for storytelling and podcasts.
The feature is free for now.
Microsoft has not mentioned any limits.
It is also unclear when it will reach mobile and desktop apps.
The MAI-Voice-1 model generates natural voices.
It creates one minute of audio in less than a second.
It was trained on 15,000 Nvidia GPUs.