FLOLO is a mindful productivity app where the AI does not suggest — it decides. Brain-dump everything on your mind, and the app returns one task to focus on right now. I built it solo end to end: UX, prompt chain, schedule logic, habit system, and the bloom that grows as work gets done.

Product — AI productivity web app

Audience — Founders, creatives, and solo operators overwhelmed by traditional task lists

Role — Solo founder, designer, prompt engineer, developer

Shipped — Live at flolo.app

Stack — Figma · Claude · Next.js · Vercel · Supabase

Why it matters — Proves an AI product can refuse to show a list and still feel like a productivity tool. The ritual is the product.

Product strategy and positioning

UX and UI built directly in Claude Code — no Figma mockup step

Brand identity and landing page

Prompt chain for task parsing + schedule reasoning

Next.js front end and Supabase back end

The bloom reward system — concept, visual, and habit integration

Beta testing and prompt tuning against real user behavior

Every user I interviewed said the same thing. They opened their task app, saw 40 items, and closed it. The list itself was the problem. The fix was not better sorting — it was refusing to show a list at all.

Instead of surfacing tasks, FLOLO hides them. The AI parses the brain-dump, reads the schedule, checks the user's energy state, and returns one sentence. "Right now, do X." Everything else stays invisible until the next check-in.

Dump — user brain-dumps messy thoughts into one field

Parse — the AI extracts tasks from unstructured text

Match — schedule and energy state filter what fits right now

Decide — the AI picks one task, not a menu

Reward — completion grows the bloom, reinforcing the ritual

Most reward systems feel tacked on. The bloom is load-bearing. It persists and grows across sessions, and users told me that visual progression is what brought them back. That reframed FLOLO: not just a productivity tool, but a ritual people return to.

The bloom is not decoration. It is a live, data-driven generative element that reads completion counts, energy state, and habit momentum in real time and generates its form from that state. I vibe-coded it end to end in Claude Code — no Figma step. To my knowledge it is one of the first artistic, state-reactive visuals shipped to production this way, pushing Claude Code past "code completer" into "design partner for generative art."

Brain-dump parser — Claude extracts tasks from messy text

Schedule-aware layer — real calendar load shapes what fits now

Energy-state selector — the AI matches task weight to user state

Habit layer — streaks and momentum reinforce the ritual

Bloom system — visual reward that grows across completed tasks

Every production prompt went through five beta loops with real users. The biggest shift was moving from "suggest a task" to "choose the task." Users didn't want options — they wanted a decision. The chain checks three signals (priority in the dump, schedule pressure, energy state) before returning a single sentence.

I vibe-coded FLOLO end to end. Claude Code and ChatGPT were the brains — every component, every flow, every interaction shipped from prompt to production with no Figma mockup step. I built a Perplexity-powered knowledge base to source user and problem research up front, then fed that context into every design and feature decision. Research → decide → ship. No handoff phase.

My Associates in Graphic Design and ongoing BFA in Electronic Design Media at CCNY are why none of this looks vibe-coded. Formal design training keeps the aesthetic tight when the build process moves fast.

I treat UX and prompt design as one system

I design product rituals, not just feature sets

I build visual reward systems that carry real product weight

I ship production AI products solo — end to end, prompt to deploy

I own the stack from research to Supabase — vibe-coded, no handoff

Research — Perplexity (user + problem knowledge base)

AI brains — Claude Code · ChatGPT · Claude

Build — Next.js · Vercel · Supabase

Brand — Pixel logo, cyan/pink palette, soft-neutral UI