Google has launched Lyria 3, a new generative AI model integrated into the Gemini application, allowing users to create original music tracks. The tool is now available in Romania, alongside a global rollout beginning in February 2026.
Lyria 3 enables users to generate unique 30-second musical pieces, complete with lyrics and custom cover art, directly within the Gemini app. The AI responds to text prompts, allowing users to specify desired styles and atmospheres. For example, a user could request “a hip-hop track with a jazz background and humorous lyrics about an alien struggling to find parking,” or create a song inspired by a specific image.
According to reports, the application can also compose music based on uploaded images or documents, analyzing the content to create a corresponding track. Users can also generate songs in Romanian, both for text prompts and lyric creation.
One user tested the application by requesting “a song that belongs to interwar Cluj, without words and based on an image from 1934, showing a horse-drawn cart in Unirii Square.”
Google has incorporated SynthID, its AI-generated content labeling solution, into Lyria 3. This allows users to verify whether an audio file was created using a Google AI tool. The model also includes filters designed to minimize the creation of music overly similar to existing works.
The launch of Lyria 3 underscores the increasing accessibility of AI-powered music creation tools, potentially democratizing the music production process for a wider audience.
Radio Romania Cluj reports that the application was made available this month.
Radio Cluj on X also announced the launch of the AI music creation tool.
ZiarulPozitiv.ro details how users can create 30-second music clips with custom covers using the new AI model.
