← BackAI Product Design · 2025

Feelink

An emotion journal that meets you where you are

Product Design · Psychology · AI-assisted build · 2025

Problem

Most mood trackers ask 'how are you?' and accept a number.

Feelink guides you through the full emotional landscape, situation, body sensation, specific emotion, trigger, because a number between 1-10 doesn't capture what you're actually feeling.

The psychology

Emotional granularity, Plutchik's wheel, and interoception

Affective neuroscience finds that emotional granularity, the ability to tell apart and label closely related emotions, predicts long-term wellbeing more reliably than how positive those emotions are. Someone who can separate frustrated from disappointed regulates better than someone who registers only bad. Plutchik's wheel gives that a structure to navigate, and research on interoception supports the second channel, that noticing where an emotion registers physically adds information a label alone does not carry.

What it changed

It is the reason the product refuses a 1 to 10 score, which is the standard mood-tracker interaction and the one that trains the opposite of granularity. Instead the taxonomy is exposed in full so the specific word is reachable, and body mapping is a first-class step. Testing changed how that step was built, not whether it existed: people only used it when the tap target was a stylised human figure rather than a list of body parts.

Start with the situation, not the score

Where are you? What happened? Context comes before the label, because emotions don't exist in a vacuum.

A taxonomy of 80+ emotions

Not just happy/sad/angry. A research-backed emotion wheel that helps you name exactly what you feel, from 'wistful' to 'exasperated'.

Patterns across check-ins

Trends emerge: which situations trigger which emotions, how your baseline shifts week to week, where your emotional vocabulary is expanding.

Flow

A guided check-in

Check-in

Situation

Taxonomy

Stack & Tools

LovableClaude (Sonnet)

design

Figma

code

React + ViteTypeScriptTailwind CSSRadix UI / shadcn

infra

CapacitorLovable hosting

Reflection

What I’d revisit

The emotion taxonomy is the strongest part. The statistics view needs more narrative: not just charts, but AI-generated summaries that say 'you tend to feel anxious in work situations on Tuesdays.'

Inner Light

More client and product design case studies at hannadevries.com