Google launches Lyria 3 and allows you to create complete songs with AI in the Gemini app

Google has launched Lyria 3, its most advanced artificial intelligence music generation model, integrated into the Gemini application. The tool allows you to create high-quality 30-second tracks from user-submitted text prompts or images. The feature has been available in beta since February 2026 for people over 18 in several languages, including Portuguese, with a focus on democratizing musical creation even for those without technical experience.

The model, developed by the Google DeepMind team, generates songs complete with vocals, original lyrics and instrumentals in a variety of genres. Usuários describe the desired style, as “upbeat pop with an electronic beat and a husky male voice”, and receive a ready-made track with a cohesive structure. Integration with photo or video uploads makes it possible to create trails that capture the atmosphere of the image, expanding creative options.

The release represents a significant evolution compared to previous versions of Lyria. Control over elements such as tempo, melodic flow, and vocal expression has improved, resulting in more natural and consistent compositions from start to finish of the track. Todas the generated tracks receive a SynthID digital watermark invisible to the human ear, but detectable by specific tools to identify content created by AI.

Access and requirements for using Lyria 3

The functionality is accessible directly in the Gemini app for web and mobile. Usuários must be over 18 years old and perform age verification to begin. The free plan offers a limited number of daily generations, while paid Google AI subscriptions expand the limits for more frequent use.

The process starts with selecting the music generation option in the app. Após enters the prompt or uploads an image, the system processes the request in seconds and delivers the banner with cover art automatically generated by another Google AI model. Creations can be downloaded in audio format or shared on social networks.

Technical improvements and creative control

Lyria 3 improved the structural uniformity of songs, avoiding abrupt transitions or deviations in rhythm and vocal tone. Usuários refine the results by specifying details such as predominant instruments, tempo or combination of hybrid genres. Exemplos include requests for pure instrumental tracks or with lyrics in multiple languages.

The ability to interpret visual prompts adds an innovative layer. A nighttime cityscape photo might generate an atmospheric electronic track, while a live concert image inspires energetic rock. Essas options encourage quick creative experimentation.

Limitations and security measures

Tracks have a maximum length of 30 seconds, making them ideal for short clips or initial songwriting ideas. The model includes filters that block prompts that attempt to imitate specific artists’ voices or play existing songs, in order to protect copyrights.

Google emphasizes that commercial use of generated lanes depends on current policies and evolving regulations. Usuários should consult the terms of service to understand legal responsibilities involved in distributing or monetizing content.

Practical application examples

Creators use the Lyria 3 to produce custom jingles, social media tracks, or prototype musical ideas. A prompt like “smooth jazz with piano and solo saxophone over nighttime reflection” results in a cohesive instrumental piece. The inclusion of automatic lyrics makes it easy to create complete songs based on personal themes.

The tool is also used for music education, allowing students to explore structures of different genres without expensive equipment. Músicos professionals test unusual combinations to inspire new compositions.

Integration with other Gemini features

Cover art generation uses the Nano Banana template to create custom visuals that complement the track. Isso facilitates immediate sharing on digital platforms. The focus on high fidelity ensures professional audio quality in short clips.

The feature aligns with Google’s strategy of expanding creative capabilities in Gemini. Future Atualizações may include longer tracks or additional controls based on user feedback.