Storio

Project Overview

No tool on the market lets you track movies, series, and books in one place while preserving the emotional context of when you experienced them, so I built one. Storio is a full-stack product I designed and developed independently from 0 to 1, completed solo using a Vibe Coding workflow from product definition and UI design all the way to deployment.

Key Achievements

• Designed an anonymous-first onboarding flow: users can add up to 10 items before signing up, letting them experience product value before committing to registration
• Built a Progressive Enrichment data model: a base Story interface for quick saves, with an extended ItemDetail interface that loads cast, streaming providers, and recommendations on demand
• Designed a three-agent AI architecture (Curator / Search / Scribe) with clear separation of concerns across content discovery, search, and AI reflection generation

Challenges

• Running Claude, Gemini, and OpenAI in parallel within the same agent architecture produced inconsistent output formats and quality, requiring a unified response schema and per-agent fallback strategy
• Next.js static export conflicted with dynamic auth states: Supabase sessions aren't accessible server-side in SSG environments, requiring a full redesign of the client-side hydration flow
• Data quality varied significantly across APIs: TMDB and Google Books differ in field completeness and naming conventions, with missing fields directly breaking UI rendering

Solutions

• Defined a unified response schema to normalize LLM outputs, with independent fallback logic per agent
• Moved all auth state management to the client side, using Supabase subscriptions to handle guest-to-user migration invisibly
• Built an adapter layer to normalize API responses using the Story interface as the minimum common denominator, filling missing fields with safe defaults

Impact

• Turned a PM's product instincts into a fully deployed, full-stack application, proving that a PM can move beyond writing specs and actually ship.
• Live and deployed, supporting Traditional Chinese and English

Tech Tags

Frontend

Next.js

React

Tailwind CSS

Backend

FastAPI

Python

Auth & DB

Supabase

Deploy

Vercel

Railway

Automation

n8n

APIs

TMDB

Google Books