Google has introduced AI-generated music capabilities to its Gemini chatbot, powered by Google DeepMind’s latest audio model, Lyria 3.
Beta access to Lyria 3 is rolling out in the Gemini app, allowing users to create 30-second music tracks directly within the chatbot using text prompts, images, or videos, News.Az reports, citing the Verge.
The feature is launching globally in multiple languages, including English, German, Spanish, French, Hindi, Japanese, Korean, and Portuguese. Access is limited to users aged 18 and older.
Lyria 3 enables users to generate songs by describing genres, moods, or personal memories — for example, requesting an Afrobeat track inspired by childhood experiences. The tool can produce both instrumental music and songs with automatically generated lyrics. Users may also upload photos or video clips, which Gemini uses to tailor lyrics and sound to match the content’s tone.
Google said the aim is not to produce “a musical masterpiece,” but to offer a creative and expressive experience. Tracks created in the app will include custom cover art generated by Nano Banana to make sharing easier.
The company is also integrating Lyria 3 into YouTube Dream Track, a feature that lets creators produce AI-generated soundtracks for Shorts.
Earlier demonstrations of Lyria and Dream Track showcased the ability to mimic well-known artists’ voices and styles. However, Google said it has taken copyright concerns seriously in developing Lyria 3. The tool is designed for original creation rather than imitation. If users request a specific artist, Gemini will generate music with a similar mood or style, while applying filters to ensure outputs do not replicate existing works.
Lyria was first introduced in 2023, but access had been restricted to Google Cloud’s Vertex platform. Expanding it to the Gemini app broadens availability, though it comes after rivals such as TikTok and Microsoft’s Copilot launched their own AI music-generation tools.