Nana banana "elevator pitch"

Gemini API
1

Hello NanaBanana Team,

I am a graphic designer, my name is Henry F. Teixeira, currently using the platform to create artwork for apparel printing, DTF, sublimation, and e-commerce products. I would like to share some technical feedback regarding the current background removal and artwork segmentation behavior.

The system works very well for:

  • single objects,

  • centered characters,

  • product photos,

  • isolated subjects.

However, professional print design workflows require a different type of composition handling:
the entire artwork composition needs to be preserved as a unified visual structure.

Currently, when working with designs containing:

  • typography separated from characters,

  • floating decorative elements,

  • logos above the main artwork,

  • paint splashes,

  • horizontal compositions,

the system often classifies those elements as “background” or “secondary objects,” removing important parts of the composition even when preservation is explicitly requested.

Based on the observed behavior, the current pipeline seems to prioritize:

  • “main subject extraction”
    or

  • “largest connected foreground object”

instead of:

  • “full composition semantic preservation”.

In practice, this creates limitations for workflows involving:

  • apparel graphics,

  • branding,

  • lettering,

  • t-shirt designs,

  • complete visual compositions.

One important example:
when typography is positioned above the main character and separated by empty space, the system tends to remove the text automatically, even when detailed instructions request full composition preservation.

Another important point:
on large print canvases (example: 4200x1000 px), the AI often:

  • minimizes typography size,

  • centers the artwork like a small sticker,

  • creates excessive empty margins,

  • or fails to properly fill the horizontal layout.

This makes professional print workflows difficult.

It would be extremely valuable to introduce features such as:

  • Full Composition Mask

  • Preserve All Foreground Elements

  • Typography Preservation Mode

  • Disable Main Subject Only Segmentation

  • Transparent PNG Print Mode

  • Wide Canvas Composition Lock

  • Sticker/Print Layout Mode

  • Connected + Disconnected Subject Preservation

These improvements would make the platform significantly more powerful for:

  • graphic designers,

  • apparel creators,

  • print artists,

  • e-commerce sellers,

  • branding professionals,

  • print-on-demand businesses,

  • visual content creators.

There is a very large community searching for tools capable of:

  • preserving complete compositions,

  • respecting typography,

  • generating true print-ready PNG files,

  • without aggressive artwork reconstruction.

I genuinely believe improving these behaviors would make NanaBanana / Genspark a very strong reference in the creative and digital printing market.

Thank you very much for your attention and for the work being developed on the platform.

Sincerely,
A graphic designer and platform user.