A digital alternative to photo shooting powered by highly accurate AI.
About
Photo AI is a creative platform by Pieter Levels that lets users generate, edit, and personalize photorealistic images and videos using artificial intelligence. The tool allows people to build custom models of themselves or others, apply different outfits, environments, and props, and even transform photos into dynamic formats such as videos or 3D assets. Its interface combines guided prompts with structured modifiers, making it possible to remix existing content, apply stylistic changes, or create entirely new media from scratch. Designed for both casual users and professionals, Photo AI streamlines what used to be complex creative workflows into a few intuitive steps, while still offering advanced controls for customization.
This project is a personal experiment where I explored ways to improve the usability and aesthetics of Photo AI. As a regular user, I initially found the interface overwhelming and fragmented, with overlapping features and unclear navigation. My redesign focused on simplifying the user flow, unifying related actions, and creating a cleaner, more intuitive sidebar structure. The goal was to reduce friction for new users while preserving the tool’s advanced creative capabilities.
The Sidebar
In my redesign, I restructured the sidebar into a clear, step-by-step flow where users first choose their input, then select the type of action (photo, remix, or video), and finally customize it with styles, props, or scenes. By unifying overlapping features and presenting them as modifiers rather than separate workflows, the sidebar feels much simpler and more intuitive. This makes it easier for users to understand what they’re creating and reduces the confusion that comes from having multiple, fragmented options.
Photo Cards
I also simplified the design of the image cards. Originally, hovering over a card revealed a cluttered set of options that competed for attention. In the redesign, I prioritized the most important actions and grouped the secondary tools under a cleaner “more options” menu. This reduces visual noise, makes the interface feel lighter, and helps users quickly find the actions they actually need without being overwhelmed.
Photo Viewer
I also redesigned the photo viewer to make it cleaner and more informative. In the original version, all actions were crammed into a horizontal bar, which made scanning difficult and pushed the photo itself into the background. In my redesign, I placed the main actions in a vertical toolbar on the right, making them easier to recognize and quicker to access. On the left, I introduced contextual information such as the prompt, model, and creation details, so users can understand the origin of each photo at a glance. This layout strikes a better balance between functionality and clarity while keeping the photo as the main focus.
Creating a Model
I redesigned the model creation flow to reduce friction and improve focus. In the original design, the process lived inside the sidebar, forcing users to scroll through long forms and making the experience feel cluttered. In my version, the flow opens in a dedicated modal, guiding users step by step. The first step asks what type of model they want to create (based on photos or an AI influencer), and the second step is where they upload images and enter additional details. This separation makes the process clearer, faster, and less overwhelming.
Occupation
Years active
Location
Age