Google has rolled out Lyria 3, a new music generation model integrated directly into Gemini, allowing users to create 30-second songs from text prompts, photos, or videos.
The feature is available starting February 18 and is launching globally for Gemini users aged 18 and older. It is initially rolling out on desktop, with mobile app access expected over the next several days.
I already tested it out, and the result shocked me.
What Lyria 3 does
Lyria 3 lets users generate short songs using simple prompts. A user can describe a theme, mood, or concept, and Gemini will produce:
- A 30-second track
- AI-generated lyrics
- AI-generated cover art (via Google’s Nano Banana image tool)
You do not need to provide your own lyrics. The model generates them automatically based on the prompt.
I asked Gemini to create a rap about a man motivated to build a better life after a breakup with a girl in the 50 Cent style.
And the results are decent:
What’s new in Lyria 3
Lyria first launched in 2023 and was later used in YouTube experiments such as Dream Track, which allowed creators to generate short background music clips for YouTube Shorts.
With Lyria 3, Google says the model now:
- Generates lyrics automatically
- Offers more control over style and vocals
- Produces more complex and realistic tracks
Dream Track is also being updated to use Lyria 3. Google has not clarified whether YouTube creators will be able to exceed the 30-second track limit.
At launch, Lyria 3 supports English, German, Spanish, French, Hindi, Japanese, Korean, and Portuguese. Additional languages are planned.
Copyright safeguards
Google says Lyria 3 is designed for original expression and not for directly mimicking existing artists.
If a user references a specific artist in a prompt, Gemini will generate a track inspired by a similar style or mood rather than copying that artist’s work.
The company says filters are in place to prevent outputs from being too close to existing songs. However, Google acknowledges that safeguards may not be perfect and encourages users to report potential copyright issues.
Google did not specify which artists, if any, contributed licensed training data for Lyria 3.
What this means for users who love music
For casual users, Lyria 3 lowers the barrier to generating short music clips. No external software or audio editing tools needed.
For creators, the current 30-second limit makes the tool most usable for short-form content, or social media posts, rather than full-length songs.
If you want to test it:
- Open Gemini on the desktop.
- Select Lyria 3 from the Tools menu.
- Enter a descriptive prompt, photo, or video.
- Review the generated track and cover art.
Not all users may see the feature immediately. Also, the tool is limited to short tracks for now, and Google has not announced plans to extend the duration past 30 seconds.
Thank you for being a Ghacks reader. The post Google Launches Gemini Music Generator With 30-Second AI Songs appeared first on gHacks.
