What is Lyria 3?

Lyria 3 is Google DeepMind’s latest high-fidelity AI music generation model, now integrated directly into the Gemini app and website. It transforms text prompts, images, or even videos into original 30-second tracks complete with instrumentals, vocals, lyrics, and custom cover art. Designed for fast, spontaneous creation, Lyria 3 lets anyone turn a funny moment, a specific feeling, or a photo into a shareable soundtrack in seconds.

Who it's for

• Casual creators – Anyone who wants to quickly generate a fun, personalized track for social sharing, group chats, or special occasions without any music production skills.
• Content enthusiasts – People who enjoy pairing photos or videos with custom music and lyrics that match the mood, vibe, or story behind their media.
• Curious experimenters – Users who love mixing genres, moods, and eras to explore new sounds and express ideas through AI-generated music.

Key features

Text and image-to-music generation

Lyria 3 turns your written prompts or uploaded photos and videos into original 30-second songs. Describe a genre, tempo, instruments, and vocal style, or simply share an image—the model interprets the context and creates a matching track with lyrics.

Built-in template gallery

Start with a curated template from the gallery to jumpstart your creation. Pick a track style, add your own details, and let Lyria 3 generate a high-fidelity soundtrack that fits your vision.

Custom cover art and sharing

Each track comes with cover art generated by Nano Banana, Google’s companion AI. You can download your song as MP3 or MP4, or share it directly to your favorite platforms from within the Gemini app.

SynthID watermarking

All Lyria 3 tracks are embedded with SynthID, Google’s imperceptible watermark for identifying AI-generated content. You can also upload an audio file and ask Gemini to check if it was created with Google AI.

What stands out

Lyria 3 turns your camera roll into a personal soundtrack studio—no instruments, no studio, just your idea and a prompt.

What sets Lyria 3 apart is its seamless integration with Gemini’s multimodal capabilities. You don’t need to describe everything in text; a photo or short video can serve as the creative spark. The model generates not only music and vocals but also matching cover art, making the output feel complete and shareable. The combination of speed (30-second tracks optimized for quick iteration) and built-in AI content identification via SynthID gives users both creative freedom and transparency.

Worth checking out if…

You enjoy experimenting with AI-generated music for fun, want to create custom soundtracks for photos and videos without any technical know-how, or need a fast way to produce shareable, personalized songs for social moments. Lyria 3 is free to use in the Gemini app and is available in all countries where Gemini is supported (users must be 18+).