Inspiration

The modern world moves fast—and our minds rarely get a break. We noticed a gap between long-form mental health solutions like therapy or meditation apps, and the need for simple, daily emotional hygiene.

We asked ourselves: What if you could mentally reset in just 5 minutes a day—with the help of AI?

Thus, WellNest was born—an AI-powered wellness companion designed to support emotional clarity, mindfulness, and personal growth through tiny, consistent habits.


What It Does

WellNest offers a “Mental Reset Capsule”—a daily 3–5 minute experience where users:

  • Journal with AI prompts (text, voice, emoji)
  • Track their mood passively
  • Get personalized affirmations or wellness tips
  • Reflect with poetic, compassionate AI replies
  • End with one small action (like breathwork or gratitude)

It’s designed to be calming, inclusive, and available in English + Hindi.


How We Built It

We used:

  • React + Tailwind CSS for a smooth, calming UI
  • Flask backend + LangChain for connecting journaling prompts and mood insights
  • OpenAI GPT-4 API + sentiment analysis models to generate responses
  • Whisper for voice journaling
  • Pinecone for storing contextual journaling embeddings
  • Google Text-to-Speech & Speech-to-Text APIs for accessibility
  • Canvas API to generate downloadable monthly “Wellness Snapshots”

Challenges We Faced

  • Passive mood tracking without wearable data was tricky—we used typing speed, word tone, and hesitation gaps heuristically.
  • Creating affirmations and reflections that feel warm, not robotic, took careful prompt engineering.
  • Making the design inclusive and accessible (voice support, local language, minimal typing) took iterative testing and feedback.
  • Handling multiple input modes (emoji, voice, text) was UX-sensitive.

What We Learned

  • The smallest habits (like 1-minute reflection) can shift emotional well-being.
  • AI can play a compassionate, non-judgmental role in mental health—if used carefully.
  • Empathy is as important in design as in the AI model response.
  • People want emotional clarity, not productivity—this shifted our approach.

🔧 Built With

  • Frontend: React, Tailwind CSS
  • Backend: Python, Flask
  • AI/LLMs: OpenAI GPT-4, LangChain
  • Voice Input: Whisper, Google Speech-to-Text
  • Sentiment Analysis: TextBlob, transformers
  • Embeddings/Memory: Pinecone
  • TTS for Accessibility: Google Text-to-Speech
  • Visual Generation: HTML5 Canvas API
  • Deployment: Render / Vercel
  • Design Tools: Figma (for mockups)

Built With

  • api
  • canvas
  • figma
  • flask-ai/llms:-openai-gpt-4
  • frontend:-react
  • google
  • google-speech-to-text-sentiment-analysis:-textblob
  • html5
  • langchain-voice-input:-whisper
  • pinecone
  • render
  • tailwind-css-backend:-python
  • text-to-speech
  • transformers
  • tts
  • vercel
  • visual
Share this project:

Updates