Google’s Lyria 3 Brings AI Music Creation to Gemini

Google has integrated Lyria 3, its latest AI music generation model, directly into the Gemini app, enabling users to create original 30-second music tracks from simple text prompts, images, or videos. This new capability, currently in beta, democratizes music creation and offers a unique avenue for creative expression.

How Lyria 3 Works

Lyria 3, developed by Google DeepMind, transforms user input into complete musical pieces, including instrumentals and lyrics. Users can describe the desired sound, genre, or mood, and Gemini will generate a track accordingly. The process is streamlined, allowing even those without musical experience to produce custom soundtracks. The accompanying cover art is generated by Nano Banana, another AI model within the Gemini ecosystem. Google’s Lyria 3 is designed to make music creation accessible to everyone.

Beyond Text: Image and Video Prompts

The versatility of Lyria 3 extends beyond text-based prompts. Users can upload images or videos, and Gemini will create a track that complements the visual content, generating lyrics that match the moment. This feature is particularly useful for creating personalized soundtracks for photos and videos. Google highlights the potential for creating tracks from everyday moments, such as turning a “welcome to the team” email into a song.

Addressing Copyright Concerns and AI Watermarking

Google has proactively addressed potential copyright issues. If a user includes a specific artist’s name in the prompt, Lyria 3 will generate music inspired by that artist’s style or mood, rather than directly copying their work. Gemini incorporates filters designed to recognize and avoid replicating existing songs. Times Now News reports on these safeguards.

To ensure transparency, all tracks created with Lyria 3 are embedded with an audio watermark using Google’s SynthID system. Gemini can likewise detect this watermark in audio recordings, verifying that the music was generated by AI. This feature is currently limited to recordings created with SynthID.

Availability and Rollout

Lyria 3 is currently available globally in English, Japanese, Korean, Hindi, Spanish, Portuguese, German, and French for users over the age of 18. The feature is rolling out first on desktop, with mobile support expected to follow shortly. Workspace Updates details the rollout schedule for Rapid Release and Scheduled Release domains, indicating a full rollout starting February 18, 2026.

Lyria 3 for YouTube Creators

Google has also released Lyria 3 to YouTube creators, following a US-only test run, allowing them to utilize the AI model to create background music for YouTube Shorts. eWeek reports on this expansion of the tool’s capabilities.

Key Takeaways

• Lyria 3 enables users to generate 30-second music tracks within the Gemini app.
• Music can be created from text prompts, images, or videos.
• Google has implemented safeguards to address copyright concerns and ensure transparency through AI watermarking.
• The feature is available in multiple languages and is rolling out globally.
• YouTube creators can now use Lyria 3 to create background music for Shorts.