Inspiration

According to WHO nearly 1 billion people are suffering from mental health problems. Due to stressful work environment, toxic people & lonelyness—they just want someone to listen. Therapy is helpful, but not always available.

To reduce that gap I created ME APP. It’s a always-available companion for emotional support—not a replacement for professionals, but a safe space between “I’m struggling” and “can I get help.”

What it does

ME is a mental health companion web app featuring Aiko, an AI companion designed to listen with empathy and care.

Key features: i) Empathetic conversations – Aiko validates feelings first, reflects emotions, and asks thoughtful questions instead of giving generic advice.

ii) Voice interaction – Natural voice responses make conversations feel more human.

iii) Calming environment – Animated ocean and aurora visuals with copyright free meditation music.

iv) Personalized support – Adapts to emotions like anxiety, sadness, loneliness, or anger using CBT-inspired techniques.

v) 3D companion – A warm, expressive anime-style avatar that feels present and approachable.

How we built it

I focused on simplicity, performance, and experience.

i) Frontend: HTML, CSS, JavaScript (lightweight, fast, responsive)

ii) 3D & visuals: Three.js with a GLB avatar and custom shaders

iii) AI: Google Gemini 2.5 Flash with carefully engineered therapeutic prompts

iv) Voice: ElevenLabs for natural text-to-speech

v) Deployment: Netlify.

Challenges we ran into

i) Making AI feel human.

ii) API limits

iii) 3D model issues.

iv) Browser audio autoplay restriction.

v) Cost control & tokens limits.

Accomplishments that we're proud of

i) Aiko feels warm, patient, and emotionally aware

ii) A calming atmosphere users actually want to stay in

iii) Fully responsive and accessible design

iv) Voice + 3D Avatar + AI in one APP.

What we learned

Audio and visuals deeply affect emotional experience

What's next for ME APP

i) Mood tracking.

ii) Guided breathing and meditation exercises

iii) Conversation memory.

iv) More voices and languages.

Built With

Share this project:

Updates