How YardRevamp works
In short: your photo is sent to an image model (Google Gemini image models, via OpenRouter) together with a fixed prompt built from the options you picked. The prompt tells the model to keep the camera angle, walls, windows, doors, the house, fence and large trees exactly as they are and to change only what you asked for. A result takes about 10 seconds.
Your photo is resized in your browser to at most 1,536 px, processed in memory and not stored on our servers. Results are returned to your browser as JPEG; we keep only anonymous counters (time, options, cost) to run the free tier.
Results are AI design previews. They can invent details, get scale slightly wrong or miss a request — generate again (failed generations are free), and always check real products, paint samples, plant hardiness and local rules before you buy or build.
Questions or a result that looks wrong? Email hello@yardrevamp.com.
Credits
Where the sample photos come from
All “before” photos are public-domain (CC0 or Public Domain Mark) images from Flickr and Openverse. The “after” images were generated with this tool, November 2026.
- backyard-japaneseR Willis · Public Domain Mark 1.0
- backyard-xeriscapeR Willis · Public Domain Mark 1.0
- backyard-patioWonderlane · CC0 1.0
- backyard-gravelJay Mantri · CC0 1.0
- front-cottageBlervis · CC0 1.0
- front-modernGlacierNPS · Public Domain Mark 1.0
- front-nativeJonathan Tasler · Public Domain Mark 1.0
- front-xeriscapeWonderlane · CC0 1.0
- patio-stonebirminghomerealty · CC0 1.0
- patio-deckJay Mantri · CC0 1.0