Tatoey

Tatoey is a personal media center for the streaming-fractured era: built by one person, for one person, with zero apology about it.

The thesis is that the real product gap is not content, it is context. You do not need another algorithm. You need something that listens to how you actually describe what you want.

Audience

Primary: One opinionated media user with fragmented streaming history and real taste context

Secondary: Anyone exhausted by platform-biased recommendation systems

My Role

UX/Product Designer and Developer - strategy, natural-language discovery, unified data layer, and implementation

Stack

React, Vite, Claude API, Trakt, TMDB, Fanart.tv

Constraint

Built entirely in GitHub Codespaces due to MDM restrictions, which forced a leaner API-first architecture

Challenge

The Problem

Streaming is fragmented. Discovery is broken. Nothing knows you.

Content is scattered across many platforms with no unified identity
Recommendation engines optimize for platform retention rather than actual taste
Watch history does not travel across services
The cognitive overhead of deciding what to watch is higher than it should be

Objectives

Project Goals

Make natural language the primary discovery surface
Use Trakt as the watch-history identity and TMDB/Fanart.tv as the metadata and visual spine
Avoid dark patterns like autoplay pressure, infinite scroll traps, and platform-biased rankings
Reduce friction between "I want to watch something" and actually watching it

In action

Feature Walkthroughs

Screens

Product Screens

Approach

Design Process

01 N=1 Strategy
- Defined single-person scope as a feature, not a limitation
- Removed social, trending, and median-user assumptions from the product model
02 Natural Language Discovery
- Made discovery start with prompts like "Something like Sicario but not as bleak"
- Used Claude to translate intent into result reasoning instead of opaque recommendation scores
03 Unified Data Layer
- Used Trakt for watch history identity, TMDB for metadata, and Fanart.tv for visual richness
- Designed the app as the aggregation layer none of those services build by themselves
04 Respectful Interaction Model
- Presented information cleanly without urgency, ranking anxiety, or autoplay pressure
- Measured interactions against whether they reduced cognitive load in the decision process

Results

Outcome & Reflection

1 Place for cross-platform watch history

AI Intent-driven discovery with actual reasoning

N=1 Built around one user's taste instead of the median

Tatoey is the product you build when every alternative has frustrated you long enough to give you real opinions.