Transforming storytelling with AI-generated content.

The Challenges

Traditional filmmaking is expensive, technical, and out of reach for most people. Plenty of people have ideas for films and series, but not the skills, equipment, or resources to turn them into something tangible.

The Solutions

I designed an interface that turns movie-making into something anyone can pick up: curating characters, locations, and styles through visual previews and simple interactive choices. The goal was to make building a custom film or series feel playful rather than technical.

Welcome to Spectra

The welcome screen features falling CDs against a vibrant backdrop, layered with a looping hero video. It's a nostalgic nod paired with a forward-looking idea of media: an entry point that sets the tone for creative exploration before the user's even signed up.

Prompt Input & Featured Content

The home screen leads with a prompt input where users can start generating a movie or series in their own words. Below it, they can scroll through content generated by others, a feed that doubles as inspiration and proof the tool actually works.

AI Video Aesthetic Builder UI Workflow

Curation starts with the visual aesthetic (realism, anime, or something more stylized), presented as a grid of options. A refresh button regenerates the set if nothing clicks, keeping exploration lightweight and fun without losing the sense of curation.

Movie/Show Description

The description screen lets users watch the selected film or episode, read a short synopsis of the story and characters, and find similar titles through recommendations tailored to what they've made. It's the layer that turns a single generated piece into a reason to keep exploring.

Generated Character Selector Screen

This Tinder-style swiping screen lets users curate their movie by swiping right on characters, styles, or locations they want, and left on what they don't. The familiar pattern lowers friction and turns customisation into something closer to a game than a form.

Generating Loader

While the movie generates, a minimal loader simply reads "Generating." It's deliberately understated, enough to keep users informed and build a little anticipation without getting in the way.

Movie Playback Screen

The playback screen uses a blurred, glassmorphic treatment behind the video for the controls and metadata to sit on. It keeps focus on the content itself while still giving the frame around it some presence.

The Process

The process started with a user persona and empathy maps to understand who I was designing for. From there I built out the information architecture, which informed the wireframes and, eventually, the final UI. Each step was in service of one thing: making AI-generated movies and series easy to actually create.

Persona

I started by mapping the needs and motivations of Spectra's users (their goals, their frustrations) into a persona: a creative person who loves experimenting with technology, wants to express themselves visually, and wants to tell stories without needing a full production studio behind them.

Empathy Maps

From there, I built an empathy map (what users said, did, thought, and felt during their creative process) to get past assumptions and design around what they actually needed. It surfaced the real motivations and pain points behind the persona.

Ideation

Next was the information architecture: organising content and defining the core user pathways so navigation felt obvious rather than learned. That structure is what let users move easily between style selection, character customisation, playback, and discovery.

Wireframes

• The information architecture became low-fidelity layouts next, focused on a clean interface for movie and series creation, with the key interactions (prompts, character selection, style customization, plot outlining) front and center. Testing flows at this stage, before high-fidelity visuals, made changes cheap.
• 1. The AI movie creation experience begins with the user writing a prompt.
• 2. Once a visual aesthetic has been selected, the user selects various aspects of the movie experience through a swiping interface.
• 3. Once all selections have been made, the movie is generated and begins playing.

Reflections

Spectra showed me what's possible when generative AI puts users in the seat of author rather than consumer. Structuring the information architecture early made every downstream decision easier, and working on a purely conceptual project taught me how much weight aesthetics have to carry when there's no working product yet: the design has to make the idea's viability feel obvious on its own.