Google Gemini Now Creates Music: Lyria 3 Enables AI-Generated Tracks with Vocals, Instruments, and Cover Art
Google Gemini's new music creation feature, powered by DeepMind's Lyria 3, generates complete 30-second tracks from text prompts — including vocals, instrumentals, and custom cover art — with SynthID watermarking.
Key Takeaways
Google has launched Lyria 3, a music creation feature within Gemini that enables users to generate complete tracks with vocals, instruments, and cover art from text prompts. The capability moves beyond simple text-to-audio to full multi-track music production.
Google has launched a music creation feature within its Gemini AI platform, enabling users to generate original music tracks from text descriptions, images, or video clips. Powered by Google DeepMind's Lyria 3 — described as the company's most advanced music generation model — the feature produces complete 30-second tracks with vocals, instrumentals, and automatically generated lyrics.
Multi-Modal Music Generation
Gemini's music creation capability goes beyond simple text-to-audio generation. Users can provide a text description ('an upbeat electronic track with female vocals'), upload an image (the AI interprets the mood and creates a matching score), or submit a video clip (Gemini composes background music that matches the visual content's tone and pacing). This multi-modal approach leverages Gemini's core strength in understanding diverse input types.
The system offers extensive customization across genre, mood, style, tempo, vocal type, and instrumentation. Each generated track comes with custom cover art, created using Google's Nano Banana technology, providing a complete package for sharing and distribution.
Technical Specifications
- Audio quality: High-fidelity 48 kHz output
- Track length: 30 seconds per generation
- Components: Vocals, instrumentals, and auto-generated lyrics
- Inputs: Text, images, or video clips
- Watermarking: Imperceptible SynthID digital watermark for AI content identification
- Availability: Global, users 18+ via Gemini app (mobile and web)
Content Integrity and Industry Impact
Every track generated through Gemini includes SynthID — Google DeepMind's imperceptible digital watermark designed to identify AI-generated content. The watermark survives common audio transformations and allows automated detection of AI-generated music, addressing one of the music industry's primary concerns about AI-created content: the ability to distinguish it from human-made works.
The feature positions Google alongside competitors like Suno and Udio in the rapidly expanding AI music generation market. However, Google's integration into the Gemini platform — which already serves hundreds of millions of users — gives the feature immediate scale that standalone music AI startups cannot match. For professional creators, Google is developing separate tools including Music AI Sandbox and MusicFX DJ, which offer greater control and the ability to generate longer or continuous music.
The music industry is watching closely. While AI-generated music creates new creative possibilities for content creators, advertisers, and casual users, it also raises fundamental questions about copyright, royalties, and the economic impact on professional musicians. How these questions are resolved will shape the relationship between AI and creative industries for years to come.