Relight, reseason,restyle.
Image-to-image AI for architecture is a class of generative models that takes an existing render and produces a transformed version — relight a daytime scene to dusk, swap summer for winter, add people and cars, change cladding material, push volumetric fog. Renovato exposes this as a node in its atlas, with thirty-plus named presets routed automatically across OpenAI (GPT Image), Google Gemini (Nano Banana), ByteDance (Seedream), and open-source models (FLUX, SDXL+).
How it works
- 01step 01 / 03
Drop a source render
Drag any architectural image — Vray, Octane, sketch, photo, or floor-plan elevation — into the atlas. It becomes a source node every preset reads from.
- 02step 02 / 03
Choose a preset
Pick from 30+ named presets — L.01 Day → Dusk, S.02 Winter snow, F.01 Volumetric fog, E.01 Add people, M.01 Material swap, and more. Each exposes its parameters as sliders.
- 03step 03 / 03
Push the variant downstream
The output becomes a node that other modes can read — animate it as a video clip, rebuild it as a 3D model, or branch it into more variants. The brief stays linked.
Presets in this mode
Each preset is named, parameterised, and routed to the engine that fits the task.
What it's for
Hero image, every season
One brochure spread needs the building in summer dusk, autumn morning, and winter afternoon. Generate all three from one source render in under a minute total.
Atmospheric variation for client review
Show the client three lighting scenarios — golden hour, overcast, blue hour — without re-rendering the model. Faster decisions, less Vray time.
Populated vs empty
Add people, cars, and street trees to a scene to convey scale and life — or strip them out for a quieter architectural statement. Two presets, one source.
Material exploration before the spec
Swap cladding from limestone to brick to bronze to weathered steel before the materials meeting. Show the partner a row of options at brief-review pace.
Frequently asked
01What is image-to-image AI for architecture?
What is image-to-image AI for architecture?
Image-to-image AI for architecture is a class of generative models that takes an existing render and produces a transformed version — for example relighting a daytime render to dusk, swapping summer for winter, adding people and cars, or changing cladding material. Renovato exposes this as a node in its atlas with 30+ named presets.
02Which AI models power Renovato's image-to-image?
Which AI models power Renovato's image-to-image?
Presets are routed automatically across OpenAI (GPT Image), Google Gemini (Nano Banana), ByteDance (Seedream), and open-source models (FLUX, SDXL+). Renovato picks the engine that best fits the preset — fast drafts on Schnell-tier open-source, photorealistic finals on Seedream, complex multi-image edits on GPT Image.
03How long does a single image-to-image run take?
How long does a single image-to-image run take?
Fast-tier presets return in roughly 3 seconds; mid-tier in 8-12 seconds; complex composition or multi-image merges in 12-18 seconds. Wall-clock time surfaces in the inspector before each run.
04Can I control the strength of the transformation?
Can I control the strength of the transformation?
Every preset exposes its parameters as sliders — strength, lighting angle, density of added elements, material colour. You can dial in the variant you want before committing credits.
05Does the output stay linked to the source render?
Does the output stay linked to the source render?
Yes. The transformed image becomes a node in the atlas with an edge back to the source. Change the source, every downstream variant updates. The Reel section on the homepage demonstrates this with six variations of one render.
06Can I chain image-to-image presets?
Can I chain image-to-image presets?
Yes. Apply L.01 Day → Dusk first, then E.01 Add people, then M.02 Brick — each becomes its own node, and the final output keeps the lineage of all three transformations.
07What does image-to-image cost in credits?
What does image-to-image cost in credits?
Fast-tier image runs are 1 credit, mid-tier 2 credits, complex composition 3 credits. Renovato's Starter tier ($19/mo) includes 1,500 credits — enough for hundreds of fast variants per month.