Let customers place a product from your catalogue true to scale on a photo of their own room and see a photorealistic image right in the browser, before they buy. It replaces the visit where goods were brought to the home just to be viewed. As an installable web app, it stays within reach for returning customers.
The Agentic Part
An image-analysis agent works out the vanishing point for every uploaded room photo and checks which wall or floor suits the placement, calibrated against a single reference point. A second agent generates the photorealistic image from that and checks the result’s quality itself; a rejected image feeds its rejection reason straight into the next attempt. Rooms already analysed stay in a cache, so a customer never has to submit the same photo twice.
Main Features
- In-browser product compositing
- Placement by dragging or corner handles
- Floor and wall placement modes
- Rotation and realistic drop shadow
- Reference-point size calibration
- Catalogue and AI service integration
- Vanishing-point and wall-suitability analysis
- Cache for analysed rooms
- Image quality scoring that feeds rejection reasons into the next attempt
- Admin-configurable AI model selection
- Installable web app (PWA)
- CI pipeline and end-to-end test suite
Use Cases
- Show a customer the product in their own room before the goods are even delivered.
- Skip the visit where goods are only brought to the home to be viewed.
- Have a rejected image feed automatically into an improved second attempt.
- Detect automatically which wall or floor suits the product placement.
What we build for you
- Product catalogue with images, dimensions and default room size per category
- Brand look and copy guidelines
- Sample room photos
- AI model choice and access keys
- Social media preview images
- Domain
- Server slot and database
The perspective calculation, the quality check and the admin area are built and only need setup.
