Angeline M
I help businesses deliver on-brand & scalable experiences
New to Contra
Angeline is building their profile!
M.Booth turns two people's photos, taken apart, into one photobooth strip,
as if they were taken together. It's built for long-distance couples who
miss the small physical proof of a shared moment, the strip from a random
Tuesday, not just the big anniversary photo.
The core mechanic: one partner shoots first, the second partner joins via
a link and poses against a live overlay of the first photo so the two line
up naturally. Three snaps each, composited into a Korean-style photobooth
strip, in a warm 90s-film style.
- Started with Claude to create prompt
- Generated design in Flowstep
- Connected to Claude Code
- Created an MVP web app : https://mbooth-nine.vercel.app/ (just the core mechanic of taking 1 photo asynchronously)
- Created images in illustrator, gen AI
- Refined, refined, refined in Flowstep and Figma
https://app.flowstep.ai/file?activeFileId=932d9289-665b-41c8-a914-d8c0f5be037a
1
1
41
Building a cohesive digital media system for Mindsets Learning
I built a cohesive brand system for Mindsets Learning, from style guide to social content, campaigns, website design, presentations, email, print materials, and video production.
1
53
One day I want to open a bagel shop. For now, I got to prototype the whole brand instead, logo, packaging, poster, video using Envato's gen AI tools. It's a surprisingly solid all-in-one kit for image and video generation.
My Workflow
- Sketched the logo on paper first, pulling inspiration from Envato's geometric patterns, I go for a monochromatic, warm, geometric logo.
- Built the final logo in Illustrator, then handed it to AI with clear sketches and direction.
- Used AI to rapid-generate mockups across t-shirts, packaging, and posters.
- Iterated through a brand animation, testing multiple storyboard approaches before landing on the final version.
- Generate music via Envato AI then edit in After Effects.
My Human by Design Moment
My first animation storyboard for the logo was low-fidelity, so the AI output felt random. I assumed the fix was more detail, so I redid it with the final logo and tighter instructions. AI overcorrected. It added motion and elements I never asked for, and the result was more confusing, not less.
I made a different call. Instead of adding more, I stripped the storyboard down to just two frames: a start and an end. No extra instruction for AI to misread, no room for it to improvise.
Designers has design in mind, our goal is to visualise that vision. AI helps along the way, but only when I gave it clarity. Simpler instruction with examples what finally got a clean, consistent animation out of it, one I could adapt across other bagel without starting over.
Curious on how other make the AI works?
Youtube: https://youtu.be/-HQX598KRK0
12
13
786