With Lyria 3, Google DeepMind brings a new level of AI music generation to the Gemini platform. Instead of complex production software or prior musical knowledge, all it takes is a simple idea – and an individual 30-second track including cover art is created within a few seconds. The integration marks a further step in the development of generative AI tools that simplify creative processes and enable new forms of expression.
- Generate AI music via text input, photo or video
- Automatically generated lyrics and individual cover art
- More control over style, tempo and vocals
- SynthID watermark for identifying AI audio
Lyria 3 in Gemini: AI music generation with creative control
With Lyria 3, Gemini is significantly expanding its creative functions. While previously the focus was on image and video generation, music is now taking center stage. The generative music model processes simple text descriptions – such as a specific genre, a mood or even an inside joke – and uses them to create a complete track. Both instrumentals and vocals are generated automatically. There is no need to write your own lyrics, as Lyria 3 develops them independently based on the prompt.
You are currently viewing a placeholder content from YouTube. To access the actual content, click the button below. Please note that doing so will share data with third-party providers.
The extended creative control is particularly relevant. Style, tempo or vocal character can be specifically influenced. This results in tracks that are not only experimental, but also sound more musically structured and realistic than previous AI models. Inspiration from visual content is also possible: photos or videos serve as a creative basis from which Gemini derives mood, context and thematic elements.
The generated songs have a length of 30 seconds and are delivered with individually designed cover art, which is created with the help of Nano Banana. The aim is not so much to produce a chart-worthy hit, but to provide a quick, entertaining way to implement personal ideas musically and share them easily. The tracks can be downloaded or shared via a link, which is particularly attractive for social media formats.
From Gemini to YouTube: Lyria 3 as a tool for creators and content
In addition to the Gemini app, Lyria 3 is also integrated into “Dream Track” on YouTube. The tool, which was originally available in the USA, is now available to creators in other countries. Especially in the context of YouTube Shorts, AI music generation opens up new possibilities for individual soundtracks. Whether short vocal passages or atmospheric background music – content can be set to music in a targeted manner without having to rely on external music libraries.
The approach to verifying AI-generated content is also technologically interesting. All tracks created with Gemini are provided with SynthID, an inaudible watermark to identify AI audio. In addition, audio files can be uploaded to the Gemini app and checked for possible AI origin. The system analyzes both the watermark and structural characteristics of the file.
This step underlines the aim of making generative AI not only more powerful, but also more transparent. Especially in the music sector, where copyright and authenticity play a major role, the traceable labeling of AI content is becoming increasingly important.
Conclusion: Lyria 3 takes AI music to a new level
Lyria 3 makes AI music generation more accessible, versatile and controllable. The combination of text, image and video input, automatic text generation and integrated verification creates a complete package for creative experimentation. The feature is now available in Gemini and via Dream Track on YouTube. Lyria 3 is thus establishing itself as an exciting tool for individual music projects – fast, flexible and technologically sophisticated.
