Gemini Expands Beyond Text with Generative Music Creation Powered by Lyria

Google continues pushing Gemini's creative boundaries, announcing today that its AI assistant can now generate original 30-second music tracks directly within the Gemini app. This expansion beyond text and image generation leverages Lyria 3 – Google DeepMind's most advanced generative music model – marking a significant evolution in consumer-facing AI tools.

The feature is rolling out in beta across Android and iOS versions of the Gemini app in eight languages: English, German, Spanish, French, Hindi, Japanese, Korean, and Portuguese. Access requires users to be at least 18 years old. While all eligible users can experiment with music creation, Google implements tiered access: Google AI Plus, Pro, and Ultra subscribers receive significantly higher usage limits compared to free users.

Generating music follows Gemini's established interaction patterns. Users describe musical concepts through text prompts (e.g., "upbeat synthwave track about summer nights" or "somber piano piece inspired by rainy days") or upload photos/videos for visual inspiration. Gemini's Lyria integration then produces both instrumental audio and automatically generated lyrics. Notably, Google emphasizes this is designed for personal expression rather than professional composition, stating the goal is "to give you a fun, unique way to express yourself."

Each generated track includes automatically created cover artwork handled by Nano Banana, Google's image generation system. Behind the scenes, SynthID – Google's watermarking technology for AI content – embeds imperceptible identifiers into the audio files. This addresses copyright concerns by enabling tracking of AI-generated origin.

Google clarifies Lyria isn't intended to replicate specific artists. If prompts reference musicians (e.g., "in the style of Taylor Swift"), Gemini interprets this as broad mood or genre inspiration rather than direct mimicry. The system also incorporates filters comparing outputs against known copyrighted material, supplemented by user reporting tools for potential rights violations.

This move places Gemini alongside specialized AI music tools like Suno and Udio, but with crucial ecosystem distinctions. Integration directly into Google's primary AI app lowers the barrier to entry significantly. The tiered access model also strategically incentivizes subscriptions to Google's AI Premium tiers, creating potential lock-in for users wanting to generate music regularly. While the 30-second limit and focus on casual creation position it differently from professional DAW integrations, it represents Google's continued expansion of Gemini from a chatbot into a multifaceted creative suite.

You can share your creations directly from the Gemini app. As generative audio evolves, tools like SynthID watermarking will be critical for transparency. Learn more about DeepMind's music generation research on their Lyria project page and explore SynthID's technical workings in Google's AI watermarking overview.