Create AI Music with Lyria 3: New Features in the Gemini App

Have you ever looked at a photo of a sunset and wished you had the perfect 30-second song to go with it? Or perhaps you’ve wanted to send a "comical R&B slow jam about a lost sock" to a friend?

Google is making that possible. In its latest innovation, Google has integrated Lyria 3, DeepMind’s most advanced music generation model, directly into the Gemini app. Currently in beta, this feature allows anyone to become a creator, turning simple text prompts or personal photos into high-quality, catchy musical tracks.

What is Lyria 3?

Lyria 3 is the next evolution of Google DeepMind’s generative audio technology. While earlier versions laid the groundwork, Lyria 3 introduces three breakthroughs that make AI music feel more human and creative:

1. Automatic Lyrics: You no longer need to be a songwriter. Gemini will generate lyrics for you based on your prompt.
2. Granular Creative Control: Users can specify the vocal style, tempo, and mood. Want a fast-paced synth-pop track or a soulful jazz ballad? Just ask.
3. Musical Complexity: The model can now handle more realistic and musically layered tracks, moving beyond simple melodies to complex harmonies.

How to Create Your First Track

The feature is designed to be playful and intuitive. There are two main ways to start composing:

1. Text-to-Track

Simply describe the vibe you’re going for.

• Example: “Create a fun Afrobeat track about memories of my mother’s home-cooked plantains.”
• Gemini will generate a 30-second song complete with lyrics and a rhythmic backing track that fits the cultural "vibe" of your request.

2. Photo/Video-to-Track

This is where the magic of "Personal Intelligence" shines. You can upload a photo,say, of your dog on a hike and Gemini will analyze the visual context to compose a song with lyrics that perfectly match the mood of the image.

Safety, Watermarking, and Responsibility

Google is acutely aware of the ethical challenges surrounding AI music. To ensure the technology is used responsibly, Lyria 3 includes several key safeguards:

• SynthID Watermarking: Every track generated by the Gemini app includes an imperceptible digital watermark. This allows anyone to verify if a piece of audio was created by Google AI.
• No Artist Mimicry: Lyria 3 is designed for original expression. If you ask it to sound like a specific famous artist, Gemini will treat the request as broad inspiration for a style rather than an attempt to clone a specific voice.
• Verification Tools: Users can now upload an audio file to the Gemini app and ask, "Was this made with Google AI?" Gemini will check for the SynthID tag and provide an answer.

Availability: Who Can Use It?

Lyria 3 is rolling out now on desktop and will reach the Gemini mobile app over the next few days.

• Who: Available to users aged 18+ in supported languages, including English, German, Spanish, French, Hindi, Japanese, Korean, and Portuguese.
• Tiers: While open to all, Google AI Premium (Plus, Pro, and Ultra) subscribers will enjoy higher daily generation limits.

The goal of Lyria 3 isn't to replace professional musicians or create the next chart-topping masterpiece. It’s about giving everyday users a new way to express themselves. Whether it's an inside joke for a group chat or a personalized birthday song, your digital life just got a whole lot more melodic.