How to Turn a Reference Photo into a Priced, Multi-Brand Specification

Clients decide with a picture — a Pinterest save, a designer's render, a magazine page. The hard part has always been turning that picture into a list of real products you can actually buy, with the right brands, sizes and prices. This guide walks through doing exactly that, step by step, so the proposal is ready before the client leaves the room.

Time required: about 5–10 minutes per project You'll need: a reference image (JPG, PNG or TIFF), and optionally a floor plan for dimension accuracy

1. Upload your reference image

Add one or more interior images. These can be renders, photos or inspiration images. Optionally include a floor plan: it helps the system infer real dimensions rather than estimating from perspective alone. No 3D models, no training, no migration — you work from the imagery you already have.

2. Let AI detect the objects

The system automatically detects and classifies the furniture, lighting, textiles and decor in the scene across 20+ categories. Review the detected objects: edit a category if needed, add anything the AI missed, and remove anything irrelevant. Every AI decision is editable — you stay in control.

3. Match each object to real products

For each detected object, the system searches across real brand catalogues and returns ranked candidates. Matching is hybrid — visual similarity plus semantic search plus business rules — not just keyword search. Apply filters to narrow results: brand, budget, category, delivery country, material and size. This is where a concept becomes a set of purchasable options across multiple brands at once.

4. Select variants

Choose the exact variant for each product where catalogue data exists: upholstery, finish, colour, leg material or size option. This is the step that turns "a sofa like this" into "this sofa, in this fabric, this size, at this price."

5. Generate the updated visual

Replace the selected objects in the original render with the chosen real products. The system preserves room geometry, perspective and lighting, so you can compare before and after and confirm the selection looks right in context.

6. Export the specification

Export a structured specification containing product images, brand and model references, dimensions, finishes, quantities, prices and room totals. Export as PDF or XLSX, with your showroom logo on professional plans. The result is a priced, multi-brand FF&E specification — the document a client says yes to.

Tips for better results

• Add a floor plan when dimensions matter; it sharpens size inference.
• Use filters early (budget and delivery country) to keep candidates relevant and avoid items the client can't actually receive.
• Keep the human in the loop — review each match before generating the visual; the AI proposes, you decide.
• Lean prompts for simple categories. For straightforward items like cabinets or wardrobes, a shorter, focused description outperforms an over-detailed one.

FAQ

Do I need 3D models or CAD files?

No. The workflow runs on 2D images — JPG, PNG or TIFF. A floor plan is optional and improves dimension accuracy.

Are the matched products real and purchasable?

Yes. Matches come from real brand catalogues with real SKUs, sizes and prices — not AI-invented products that exist nowhere.

Can one specification include products from several brands?

Yes. The entire point is assembling a cross-brand proposal — multiple brands in one scene and one priced specification.

What formats can I export?

PDF and XLSX, including product images, SKUs, dimensions, finishes, prices and room totals. Professional plans add your showroom logo to exports.