Clone and deploy your own AI interior design app that shows users their room in a new style.
Use RoomGPT as a starter to build a niche redesign app for offices, retail spaces, or outdoor areas.
Launch your own copy on Vercel in minutes with the one-click deploy button to experiment with ControlNet image generation.
Learn how to integrate Replicate's AI model API and Bytescale file uploads into a Next.js application.
Requires a Replicate account and API key, optional Upstash Redis for rate limiting.
RoomGPT is a web app that lets you upload a photo of a room and receive an AI-generated redesign of that same space. You take a picture of your bedroom, living room, or any other room, and the app returns a new version of it in a different style. Under the hood, the app sends your uploaded photo through a machine learning model called ControlNet, which generates a variation of the room while preserving its basic structure and layout. The model runs on Replicate, a platform for hosting AI models via API. Image files are stored using Bytescale. The web app itself is built with Next.js, a JavaScript framework for building websites. This repository is the original open-source version of RoomGPT, released before the paid SaaS product at roomGPT.io was built. It does not include user accounts, payment processing, or the additional features added to the commercial version. According to the README, it is intentionally simple so that developers can clone it, set it up quickly, and experiment. To run it yourself, you create an account on Replicate to get an API key, add that key to a local environment file, install the dependencies, and start the development server. There is also a one-click deploy button for Vercel, which lets you publish your own copy of the app to the web without setting up a server manually. Optional rate limiting can be added by connecting a Redis database through Upstash. The code is released under the MIT license.
← nutlope on gitmyhub — every repo by this author, as a profile.
Verify against the repo before relying on details.