Toxic Positivity Breaker

Replaces AI-generated toxic positivity with raw human empathy. Highlight 'Good vibes only!' becomes 'It's okay not to be okay'. Transforming dismissive platitudes into emotionally validating ones.

Comment

Inspiration

Watching friends absorb harmful "positive" slogans during hard times. We realized AI-generated perfectionism erodes authentic connection – this fights algorithmic invalidation at its source.

What it does

Detects 20+ toxic positivity phrases (e.g., "Just stay positive!") and replaces them with human-curated empathetic responses (e.g., "Your feelings are valid") in real-time. Includes visual highlighting and toggle.

How we built it

  • Frontend: Vanilla JS DOM manipulation with MutationObserver
  • Chrome API: Manifest V3 for content scripts
  • Phrase Database: 15 pre-curated replacements
  • Visuals: CSS glow effects + emoji transformations (💔→❤️‍🩹)

Challenges we ran into

  • Avoiding infinite replacement loops in text nodes
  • Handling React/Angular SPAs without breaking functionality
  • Performance on heavy sites like Facebook
  • Creating non-intrusive visual cues

Accomplishments we're proud of

  • Zero false positives in testing
  • 200ms average replacement speed
  • Personalizing responses with our team's authentic voice
  • Solving a real pain point in <4 hours

What we learned

  • Chrome's content security policy limits
  • Text node traversal efficiency
  • Linguistic patterns of toxic positivity
  • Micro-interventions' mental health impact

What's next for EmpathEase

  1. Crowdsourced phrase database + voting
  2. Context-aware replacements (grief vs. frustration)
  3. Firefox/Safari ports
  4. "Teach Mode" for custom phrases

Built With

Share this project:

Updates