Lyria 3 turns written descriptions into 30-second music clips without any music production skills. If you need background music for a video, a quick jingle for a social post, or a mood-setting track for a presentation, you type what you want and the model handles the composition. It responds to both text prompts and image inputs, so you can feed it a photo of a rainy city street and it will interpret the visual mood as audio. The model reads prompts that specify genre, instruments, tempo, and emotional tone, so the output matches what you had in mind rather than a generic loop. You can pair an image alongside your text to add a visual dimension to the musical direction. Each generation produces a polished clip that works straight out of the box for social content, game audio, or podcast intros. Lyria 3 fits into any content workflow where original audio would otherwise require a composer or a stock music subscription. Drop it into a video editing session, attach the clip to a social post, or use the output as a starting point for further production. Try it now and go from idea to finished track in under a minute.
Lyria 3 turns a short text description into a 30-second music clip, no instruments, no studio, no music theory required. If you need background audio for a video, a custom jingle for a brand pitch, or a mood-setting loop for a game prototype, you describe what you want and the model composes it. Available on Picasso IA, it runs entirely in the browser with no software to install. You can also feed it up to 10 images and let the model interpret the visual mood into sound, which helps when you have a scene reference but struggle to put the music into words.
Do I need programming skills or technical knowledge to use this? No, just open Lyria 3 on Picasso IA, adjust the settings you want, and hit generate.
Is it free to try? Yes, you can run Lyria 3 directly in the browser without a paid subscription to get started. Check the current pricing details on the model page for information on generation credits.
How long does it take to get results? Most generations complete within seconds. The exact time depends on server load, but you typically have a playable clip within 15 to 30 seconds of submitting your prompt.
What output formats are supported? Lyria 3 returns an audio file you can play back immediately in the browser. Download it and use it in video editors, game engines, or any audio software that accepts standard audio formats.
Can I use images instead of a text prompt? Yes. Upload up to 10 images as a visual reference and the model will interpret the mood, colors, and composition to shape the music. Combining images with a text prompt gives you more precise control over the result.
Where can I use the outputs? The generated clips work in personal projects, demos, prototypes, and creative work. Review the usage terms on Picasso IA for commercial licensing details before publishing.
What happens if I am not happy with the result? Refine your prompt with more detail about the instruments, tempo, or mood you want, then regenerate. Small changes in wording often produce noticeably different results, so iteration is a normal part of the process.
Everything this model can do for you
Type a description of the genre, instruments, mood, and tempo to receive a 30-second original clip.
Upload up to 10 reference images and the model translates visual mood into a matching musical composition.
Receive a finished audio clip in seconds without waiting for lengthy rendering queues.
Describe what you want in plain language and the model handles all compositional decisions.
Generated clips are original compositions, not sampled loops, so you can use them in published content.
Specify BPM, scale, instruments, and emotional tone for output that closely matches your creative brief.
Combine a text prompt with one or several images for a more specific and layered creative direction.
