There’s a new way to soundtrack your life — and it only takes a prompt.
On 18 February, Gemini introduced custom music generation in its app, powered by Lyria 3, Google DeepMind’s latest generative music model. Rolling out in beta, the feature allows users to create 30-second original tracks simply by typing a prompt or uploading an image.
Yes, that means your “comical R&B slow jam about a sock finding their match” can now exist in real life.
From text to track in seconds
With Lyria 3, users can describe a genre, mood, memory, or even an inside joke — and Gemini will transform it into a fully produced 30-second song, complete with generated lyrics if you want them.
Feeling nostalgic? You could prompt: “Create a fun afrobeat track about my mother’s home-cooked plantains and our childhood memories.” Within seconds, you’ll get a custom audio clip ready to download or share.
No lyrics? No problem. Lyria 3 generates them automatically based on your idea.
More control, more complexity
According to Gemini, Lyria 3 advances its previous music models in three key ways:
-
Auto-generated lyrics based on your prompt
-
Greater creative control over style, vocals, and tempo
-
More realistic and musically complex tracks
Users can also upload photos or videos to inspire a track. For instance, share a photo of your dog on a hike, and Gemini will compose a song that matches the mood — lyrics and all.
Each 30-second track comes with custom cover art generated by Nano Banana, making it easy to share directly via download or link.
The goal isn’t to produce your next chart-topping single, but to give fans a fun, expressive way to create their own mini soundtrack.
Shorts creators, take note
Beyond the Gemini app, Lyria 3 is also powering enhancements on YouTube’s Dream Track for Shorts creators in the U.S., with expansion to other countries underway.
From lyrical verses to vibey backing tracks, creators will be able to customise soundtracks for their Shorts more precisely — levelling up the overall quality of short-form content.
Built with guardrails
All music generated in the Gemini app is embedded with SynthID, Google’s imperceptible watermark that identifies AI-generated content. The app also allows users to upload audio files to check if they were created using Google AI.
Gemini says Lyria 3 was developed in collaboration with the music community, with attention to copyright and partner agreements. The tool is designed for original expression — not to replicate specific artists. If a user names an existing artist in a prompt, Gemini interprets it as broad inspiration in style or mood, not imitation.
Users are also required to adhere to Gemini’s Terms of Service and generative AI policies.
Who can use it?
Music generation with Lyria 3 is available to Gemini app users aged 18 and above in English, German, Spanish, French, Hindi, Japanese, Korean, and Portuguese. It’s rolling out on desktop first, with mobile access arriving over the next several days.
Google AI Plus, Pro, and Ultra subscribers will receive higher usage limits.
For fans who’ve always wanted to turn a memory, meme, or mood into a song — your 30-second soundtrack might just be a prompt away. Try it out today at gemini.google.com.
Like what you read? Show our writer some love!
-