Image generation API for architecture
Describe a building, a room, a garden or a product and get a photoreal image back, with no source image needed. The same engines that render from sketches and photos in mnml Studio.
From words alone
Leave the image out and the render runs from the prompt, in text-to-render mode.
Steer it with references
Add images whose style, materials, mood, colour or geometry the result should follow.
Any frame
Pick the frame of the result, from a square to a cinematic 21:9.
Variations in one call
Ask for up to 4 takes on one prompt; each is its own job and charge.
Start with a single call
Every call returns a job. Poll it, wait on it, or add a webhook and hear when it settles. The SDKs for TypeScript, Python, PHP, Go and .NET send the same request.
- v4.6$0.2475 per image
curl -X POST https://api.mnml.ai/v2/renders \
-H "Authorization: Bearer $MNML_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"engine": "v4.6-ultra",
"prompt": "A three-storey timber house on a hillside, floor-to-ceiling glass, golden hour",
"aspect_ratio": "16:9",
"count": 2
}'- Mood images and concept options inside a design or briefing tool
- Room and product imagery for a furniture or materials catalogue
- Visuals for property marketing before anything is built
Questions
Do I need to send an image?
No. A render without an image runs from the prompt alone. Add an image when you want the result to follow an existing design.
What can it generate?
Exteriors, interiors, masterplans, landscapes, plans and products: the subjects mnml’s engines are built for.
How do I get several options?
Set count from 1 to 4. Each variation is its own job and charge, listed in the response.
How is it priced?
Prepaid, pay as you go. Top up any amount from $10 to $5,000 and each call draws its listed price from your API wallet. Your first top-up adds $5 free. There is no subscription.
Start rendering from your own product.
Top up from $10, and your first top-up adds $5 free. The same account signs you in to Studio and the API.