Upload → Describe → Design → Transform
An intelligent interior design assistant that converts any room photo into a fully structured, purpose-driven design brief — built by Ayush Nandi.
▶ Live DemoRoomRevamp AI is a front-end interior design tool that lets users upload a photograph of any room, specify the room's intended purpose and vibe, and instantly receive a structured design brief — covering style direction, colour palette, furniture layout, lighting strategy, decor accents, and budget-tiered shopping priorities.
The prototype is fully functional as a standalone UI, and is architected to plug directly into any vision/image AI model for full end-to-end AI analysis in production.
Drag-and-drop or click-to-browse room image upload. Supports JPG, PNG, and WebP with instant live preview.
Choose from Bedroom, Living Room, Kitchen, Home Office, Studio, Kids Room, or Balcony to steer the design direction.
Get tailored recommendations across Budget-friendly, Mid-range, and Premium spending levels — practical for every situation.
Describe the purpose, vibe, constraints, and mood in plain language for a fully personalized brief.
Generates a complete plan: style direction, colour palette, furniture placement, lighting, decor, and shopping priorities.
Fully mobile-friendly layout, optimised for all screen sizes from phones to widescreen monitors.
Drop a clear photo of the existing space. The app previews the image instantly, giving you a before-state reference while you craft your brief.
Select the room type and budget level, then write a free-form description — the intended purpose, preferred style, specific constraints, and any must-haves you want to keep.
Hit Generate and receive a structured design brief you can act on immediately — or pipe directly into an AI image model for a visual reimagining of the space.
Every suggestion ties directly to the declared room function. A gaming setup gets different lighting and layout advice than a nursery — the brief is always contextually grounded.
Recommendations are calibrated to three distinct budget tiers, ensuring the output is actionable rather than aspirational. Users can see what's possible at every level.
Designed as a front-end prototype that seamlessly connects to any vision/image AI model via API. The interface is complete — plug in GPT-4o Vision, Claude, or Gemini and it's production-ready.
Three steps, zero accounts, zero installs. Upload, describe, generate. The fastest path from "I need to redesign my room" to an actionable plan.
Drag-and-drop or click upload with instant room preview.
Contextual selectors that steer the entire design brief.
Structured output from natural language intent.
Connect to GPT-4o Vision or Claude for real AI-powered room analysis.
Generate photorealistic room reimaginings via DALL-E or Stable Diffusion.
Direct links to recommended furniture and decor items on major retailers.
Contributions, issues, and feature requests are welcome! Feel free to check the issues page.
This project is open-source under the MIT License.