Interio is a low-code web application built to ease the process of interior design visualization. The platform’s main function was processing an image of a room and returning a visualization in a style chosen by the user. The process involved connecting to a Multi-ControlNet AI model through Replicate’s API, which supported both image generation and inpainting. One of the features included an input field that enabled users to request specific modifications from the AI model. Additionally, the interactive canvas allowed users to draw directly on the generated images, highlighting specific areas where changes were needed.
Challenges
The primary obstacle was integrating an AI model capable of creating aesthetically pleasing images with minimal visual artifacts. The solution had to support a subscription-based payment model with a credits system. Additionally, it was essential to develop an intuitive design that ensured seamless user-friendly experience.
Solutions
The main steps in developing a solution involved: selecting the appropriate AI model, fine-tuning parameters to minimize visual errors and enhance cohesion, and determining which parameters should remain adjustable for the user. It featured a custom-coded HTML canvas that allowed users to adjust brush width and select colors — black to exclude and white to include specific areas of an image for inpainting. The payment system was implemented using Stripe’s API, which handled subscriptions and credit purchases.
Outcome
Within first months of the platform’s public release, the users generated over 1 000 visualizations. The clients’ feedback led to improvements to the AI model that greatly improved the quality of the images.