Google has introduced a new music-generation feature inside the Gemini app, allowing users to create short songs using text prompts, photos, or videos. The feature is powered by Google’s Lyria 3 model and is currently available in beta.
With this update, users can describe the mood, genre, or theme of a song, and Gemini will generate a 30-second track along with matching cover art created using Nano Banana. Tracks can be downloaded or shared via a link, making it suitable for quick creative expression rather than professional music production.
According to Google, Lyria 3 builds on earlier internal music models with improvements in control and flexibility. The company said that users do not need to write their own lyrics, as the system can automatically generate them based on the prompt.
Google stated that the model provides more granular control over style, vocals, and tempo. The app generates a 30-second song along with AI-created cover artwork.
Google outlined two primary methods for using the feature:
Users can enter a detailed description of the song they want — including genre, mood, personal memory, or even an inside joke. Gemini then generates a complete short track based on that prompt.
Users can upload a photo or video. Gemini analyzes the content and creates a song with lyrics and tone that match the visual mood.
This multimodal approach reflects Google’s broader push into combining text, image, video, and audio generation within a single AI ecosystem.
Google also confirmed that Lyria 3 can be accessed through Dream Track on YouTube.
Dream Track is a suite of experimental tools designed to help creators generate unique music for YouTube Shorts. It produces AI-generated soundtracks based on text prompts.
Currently available in the United States, Dream Track is gradually expanding to creators in other regions.
Google noted that YouTube Shorts creators can now generate higher-quality soundtracks using Lyria 3, including short songs with lyrics or background music aligned with their video’s mood.
Google said that all tracks generated in Gemini will include SynthID, a digital watermark embedded into the audio to identify AI-generated content. The watermark is not noticeable to listeners.
In addition, Google is expanding Gemini’s verification feature to audio files. Users can upload a file and ask whether it was created using Google AI. Gemini will check for the SynthID watermark and analyze the content before responding.
This builds on Google’s broader AI transparency efforts across images and videos.
Google emphasized that Lyria 3 is designed to support original creation and not to replicate specific artists.
If a user references an artist in a prompt, Gemini treats it as stylistic inspiration rather than direct imitation. The system generates music with a similar mood or style but avoids copying identifiable elements.
The company also stated that filters are in place to compare outputs against existing content. Users can report content if they believe it violates intellectual property rights.
Lyria 3 is now available in the Gemini app for users aged 18 and above in the following languages:
English
German
Spanish
French
Hindi
Japanese
Korean
Portuguese
Google said support for additional languages will be added over time.
The feature is rolling out first on desktop and will arrive on mobile devices in the coming days. Google AI Plus, Pro, and Ultra subscribers will receive higher usage limits.
Google’s integration of Lyria 3 into Gemini signals a major step in AI-powered creative tools. By enabling multimodal music generation — from text, images, and video — the company is positioning Gemini as a comprehensive creative assistant.
At the same time, the inclusion of SynthID watermarking and artist safeguards reflects the industry’s growing focus on responsible AI deployment.
As AI music tools continue to evolve, Gemini’s new feature highlights how generative AI is moving beyond text and images into more immersive creative experiences.