Nano Banana Image Generator
The Nano Banana adapter is a slot in the image engine registry. It is not enabled on this deployment — here is what it does once an operator adds the key, and what you can generate right now.
Image generation runs through pluggable engines, exactly like the 3D studio. This deployment ships a local procedural engine that always works and produces real images. External image engines activate only when their API key is configured.
External image engines are not enabled on this deployment. No external image engine is configured, so generation runs on the built-in procedural engine. It produces real PNG output — patterns, gradients and surface textures — driven by your prompt.
Prompt
Nano Banana is an external image model reached over its own API: you send a prompt and, optionally, reference images, and it returns a generated or edited picture. On this site it is wired in the way every engine is wired in — as an adapter behind one interface — so the prompt box, the style and size controls and the download path do not change when an engine is added or swapped.
Status on this deployment: not enabled. No API key is configured, the adapter reports itself unavailable, and the engine picker shows it marked as not configured next to the local procedural engine. The registry refuses it with an engine_unavailable error rather than returning a picture. Every image this site produces today comes from the local engine, and nothing on this page claims otherwise.
How it works
Check what is available before you ask for it
The engine picker lists every image engine with its state. The local procedural engine is available; the Nano Banana adapter is listed as not configured. Two seconds of reading saves a wasted generation.
Generate with the engine that is available
The local engine renders patterns, gradients, terrain, textures and blueprint drawings from your prompt. Same prompt and settings, same bytes — no key, no queue, no third party in the path.
To enable the adapter, as the operator
Enabling an external engine is a deployment task rather than a user setting: the vendor's API credentials go into the configuration, and the adapter reports itself available, which switches the picker entry on for everyone. Until that happens it stays off, and this page keeps saying so.
Nano Banana Image Generator
One interface, several engines
The adapter implements the same generate() contract as the local engine: a prompt, a style, a size and an optional seed in; PNG bytes, a palette and a report out. Swapping engines does not change how the tool is used.
Availability you can see
Every engine publishes a name, a label, a note and an available flag. The picker shows the flag, so 'not configured' is visible before you click rather than after.
A clear refusal instead of a silent fallback
When an engine is not configured, the registry raises engine_unavailable with the reason attached. Nothing is charged for a request that cannot run, and no placeholder image is passed off as a result.
The local engine still answers
An unconfigured adapter never leaves the site without an image generator. The procedural engine is always available and handles the prompt on its own.
No lock-in by design
Because the adapter is one module behind a protocol, an operator can add, swap or remove an external image engine without touching the tool pages or the content that describes them.
Image tool questions
Is Nano Banana running on this site?
No. The adapter slot exists in the engine registry, but no API key is configured for it on this deployment, so it reports unavailable. Nothing on this site returns a Nano Banana image today.
Can I enable it myself?
No. Enabling an external engine is a deployment task that needs the vendor's API credentials in the configuration. There is no user-facing switch, and the picker will not offer a key field.
What happens if I ask for it anyway?
The registry refuses with an engine_unavailable error that names the adapter and says a key is required. You get a clear message instead of a broken image or a charge.
What can I use in the meantime?
The local procedural engine on the AI image generator page: five styles — pattern, gradient, terrain, texture and blueprint — four canvas sizes, and byte-identical results for the same prompt.
Why list an engine that is not available?
Because hiding it would be worse. The engine picker is the honest inventory of what this deployment can run: one engine that works, one adapter waiting for a key. The 3D studio handles its remote engines the same way.
Does my prompt get sent to a third party?
Not today. The local engine renders on this server. If an operator enables an external engine, that engine would receive the prompt and possibly a reference image — which is exactly why the picker names the engine before the request is sent.
What would the adapter do once configured?
Send the prompt and any reference images to the external model through its API, return the resulting image as PNG bytes, and report it in the same result shape as a local render: same palette field, same report, same download path.
Is this page a product page for Nano Banana?
No. It documents an adapter in this site's engine registry: what it is, what it would do, and that it is not enabled here. The model itself is a third-party product with its own terms and its own pricing.