October 2026 so far: Music and more creative choices

We are only three days into October, and the studio has already grown in several directions. Music has its own workspace, and image and voice creators have more choices to explore.

We are also making it easier to understand each tool before you commit to using it. More choices are useful when you can see what they are for and get to the relevant controls without losing your place.

Give a project a soundtrack

Music now has a dedicated composer with Google Lyria 3, Lyria 3 Pro, and ElevenLabs Music. You can describe a direction, adjust the options the selected model supports, listen to the result, and download it. Music joins the same saving, wallet, Gallery, and feedback flow as other audio.

For a short travel montage, try describing a gentle instrumental track with acoustic guitar, a steady pace, and a hopeful mood. For a product demo, you might want something restrained that leaves room for the narration.

Some models support duration controls or optional image inspiration. The controls depend on the selected model, so the studio shows the relevant options rather than expecting every music model to behave the same way.

More ways to work with an image

New image choices include FLUX 3 Image, Qwen Image 3 and Pro, layered-image output, multi-angle editing, and Grok image variants. The layered workflow keeps the layers returned by the model, making it more useful when you want to inspect the separate pieces of a composition.

A multi-angle option can be worth exploring when one view of an object is not enough for a concept presentation. A layered result can give you a different way to think about a layout. These are model-specific approaches, so start with the task you have rather than assuming you need every available setting.

A broader voice studio

Voiceover now brings more of the speech catalogue into one workspace, with the right voice, language, delivery, and output controls for the selected model. Gemini's sample voices remain available where provided, and reusable speaker settings help keep a direction handy.

If you are testing several episodes of the same explainer series, a saved speaker setup gives you a consistent starting point. You can still adjust the delivery for a different script instead of rebuilding the whole choice from scratch.

Better ways to find the tool and tell us what is missing

The public catalogue can open a selected model in the right workspace. Sign-in keeps that destination, and the welcome step lets you continue exploring or add wallet credit. Shared settings panels, clearer model rows, helpful tooltips, and category colours make the studio easier to navigate.

There is also a contact form in Settings for feedback, bug reports, feature requests, and other enquiries, with screenshot attachments. A short description of what you were trying to do and what happened is useful context for the next improvement.

October is still underway. This update covers the publicly available creative tools developed through 3 October 2026.