Image Tools
Generate an image from a prompt in five styles, download the PNG, or hand the same file to the 3D studio as a reference image.
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.
AI Image Generator
Type a prompt, pick a style and a canvas size, and get a real PNG back in well under a second. Rendered on this server — no API key, no queue.
AI Pose Generator
Re-posing a photograph is image editing and needs an external engine key. Generating a posed 3D figure works today on the local engine — this page is clear about which is which.
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.
How these tools work
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.
| Engine | Status | Notes |
|---|---|---|
| Local Image Engine procedural-image |
Available | Runs on this server. Draws patterns, gradients, terrain, textures and blueprint drawings from the prompt's seed. Always available, no API key needed. |
| Nano Banana (external adapter) nano-banana |
Needs API key | Adapter slot for an external image-editing model. Not enabled on this deployment: it needs an API key before it can be selected. |
Image tool questions
What can I do with these image tools?
Generate an image from a text prompt, choose the style and the canvas size that suit the job, download the PNG — and when you want more than a picture, send the same file into the 3D studio as the reference for image-to-3D.
Which image engines run on this site?
Image generation is a pluggable engine slot, exactly like the 3D studio. The local procedural engine is built in and always available; external image engines are adapters that switch on only when the operator configures their API key. No external image engine is configured on this deployment, so every render here comes from the local engine.
What does the local engine actually produce?
Real pixels computed from your prompt: geometric patterns, multi-stop gradients, noise-based terrain with hillshading, tileable surface textures and blueprint-style technical drawings. Each render is a genuine PNG file — no stock library, no placeholder image.
Is image generation deterministic?
Yes. The prompt is hashed into a 32-bit seed that drives every random decision, so the same prompt with the same style and size returns byte-identical PNG files — today, tomorrow, and after a restart.
What sizes can I generate?
Four: square at 768×768, landscape at 1024×576, portrait at 576×1024, and wide at 1280×512.
Can I use a generated image as a texture on a 3D model?
The texture style is built to tile: its lattice wraps, so the left edge matches the right and the top matches the bottom, and the repeat disappears when the map is applied. Image-to-3D also reads the reference image's colour, contrast and aspect ratio, so the colours you generate carry into the model.
Does my prompt get sent to a third party?
Not on this deployment. The local engine renders on this server and your prompt never leaves it. If an operator ever enables an external image engine with an API key, that engine would receive the prompt and any reference image — and the engine picker names the engine before the request is sent.
Which of the three pages do I need?
The AI image generator for prompt-to-PNG in any of the five styles. The AI pose generator if you are trying to get a person into a different pose: it sets out which route works here and which one needs an engine key. The Nano Banana page documents that adapter's status — the slot exists, it is not enabled on this deployment, and the page says so.