Google has officially added music generation to the Gemini app, giving everyday users the ability to conjure original songs from a simple text prompt with no instrument required.
The feature is powered by Lyria 3, the latest generative music model from Google DeepMind, and is currently rolling out in beta. To use it, you describe the song you want, and Gemini generates a track complete with lyrics. The classic example Google offers, a “comical R&B slow jam about a sock finding its match”, gets the point across. The output is a 30-second track with cover art included.
More Than Just a Prompt Box
Where Lyria 3 distinguishes itself from earlier iterations is in both the depth of creative control and the quality of what it produces. Users have more control over elements like style, vocals, and tempo, and can create more realistic and musically complex tracks, with lyrics generated automatically from the prompt, no input required.
There’s also a multimodal angle: you can upload a photo or video, and the tool will generate a song to match the mood of the media. It’s a feature that could find a natural home in social media content creation, personal projects, or just goofing around.
The Artist Question
Predictably, the arrival of a consumer-grade AI music tool raises questions about what it means for artists. Google has tried to pre-empt concerns by drawing a clear line between inspiration and imitation. While you can include an artist’s name in your prompt, Gemini will treat it as broad creative inspiration, generating a track in a similar style or mood rather than mimicking the artist directly.
Whether that distinction satisfies the music industry remains to be seen. AI music companies are currently navigating a minefield of copyright lawsuits, and Google’s careful framing suggests it’s aware of the legal and reputational stakes involved.
Watermarks and Wider Access
On the transparency front, all tracks generated through Lyria 3 are embedded with SynthID, Google’s watermark for AI-generated content. Google is also expanding SynthID’s detection capabilities within Gemini itself, so users can upload a track and ask whether it was made by AI — a neat bit of self-awareness baked into the product.
The feature is available to all Gemini users aged 18 and over, with support for English, German, Spanish, French, Hindi, Japanese, Korean, and Portuguese. Higher usage limits are reserved for paid subscribers on Plus, Pro, and Ultra plans. Lyria 3 is also rolling out to YouTube creators globally through the Dream Track feature, expanding a tool that was previously limited to U.S. users.
A Creative Suite Taking Shape
With image generation, video creation, and now music all living under the same roof, Gemini is starting to look less like a chatbot and more like a full creative suite. The music feature is still in beta, which means rough edges are to be expected, but the ambition is clear. Google wants Gemini to be the place you go when you want to make something.
