Google Lyria 3: Gemini App Can Now Create 30-Second Music Tracks From Text or Photos

Google just made the Gemini app a music studio. Lyria 3, Google DeepMind's most advanced music generation model, is now rolling out in beta in the Gemini app — and it lets anyone create custom 30-second tracks just by typing a description or uploading a photo.
Official Short
Watch It in Action
What's New in Lyria 3
Lyria 3 improves on previous versions in three key ways:
- Auto-generated lyrics — you don't need to write any words; the model generates them from your prompt
- More creative control — you can specify the style, vocals, and tempo you want
- More realistic and complex tracks — higher quality audio with stronger musical structure
How to Use It
There are two main ways to create tracks in the Gemini app:
- Text to track — describe a genre, mood, inside joke, or memory. Example: "I'm feeling nostalgic. Create a track for my mother about the great times we had as kids and the memories of her home cooked plantains. Make it a fun afrobeat track with a true African vibe."
- Photo/video to track — upload a photo or video and Gemini composes a matching track with lyrics. Example: "Use these photos to create a track about my dog Duncan on a hike in the woods."
Each track comes with custom cover art generated by Nano Banana and is easy to share via a download link.
Also on YouTube: Dream Track
Lyria 3 is also powering YouTube's Dream Track feature, which lets creators generate unique Shorts soundtracks — either a lyrical verse or a backing track — to pair with their videos. Dream Track is available in the U.S. and now rolling out to creators in other countries.
SynthID Watermarking & AI Verification
All tracks created with Lyria 3 are automatically embedded with SynthID, Google DeepMind's imperceptible watermark for AI-generated audio. You can now also upload any audio file to Gemini and ask if it was made with Google AI — Gemini will check for SynthID and reason about the result. This expands Gemini's verification capabilities beyond images and video to now include audio.
Copyright and Responsible Design
Google has been developing Lyria in collaboration with the music community since 2023. A few key guardrails:
- Lyria 3 is designed for original expression, not mimicking specific artists
- If you name an artist, Gemini takes it as broad creative inspiration (similar style or mood) rather than a direct imitation
- Filters check outputs against existing content to flag potential matches
- Users must follow Google's Terms of Service and Gen AI prohibited use policies
Availability
Lyria 3 is available now in the Gemini app for users 18 and older in:
- English, German, Spanish, French, Hindi, Japanese, Korean and Portuguese
- Rolling out on desktop today, with mobile app access over the next several days
- Google AI Plus, Pro and Ultra subscribers get higher usage limits
Try it at gemini.google.com/music.
The Bottom Line
Lyria 3 isn't aimed at replacing professional music producers — it's a fast, fun way to add a custom soundtrack to any moment. The combination of text-to-track, image-to-track, auto lyrics, and SynthID watermarking makes it one of the more thoughtfully built generative audio tools to launch this year. If you've ever wanted background music for a memory, a joke, or a pet video, this is worth trying.