MannSahay | Devpost

home page
links to directly access pages
friend bot

Inspiration

MannSahay was inspired by a simple but painful truth: today’s youth is surrounded by people, screens, and noise—yet feels deeply alone.
We saw friends, classmates, and even ourselves struggle silently with homesickness, academic pressure, language-based mockery, and the fear of being judged for “overreacting.”

Statistics made it worse. Suicide being a leading cause of death among Indian youth was not just a number—it was a warning.
The biggest gap wasn’t stress, it was access: long waitlists, stigma, language barriers, and a severe shortage of mental health professionals. We asked ourselves: If help cannot reach everyone, can technology become the first bridge?} That question became MannSahay.

What it does

MannSahay is a 24/7 emotional support companion designed specifically for students and young adults.

It offers:

MannSahay (Guide Bot): helps users identify, name, and understand their emotions with gentle, psychologist-guided conversations.
MannMitra (Friend Bot): a warm, casual, peer-like companion that listens without judgment and provides emotional relief through companionship.
Incognito Mode: zero logs, zero history—thoughts disappear once the chat ends.
Chat With Anyone: real-time anonymous peer chats for human-to-human connection.
Activity Zone: journaling, gratitude, breathing, soundscapes, and mini relaxation games.
Offline Mode: emotional support even without internet using local LLMs.

It’s not therapy.
It’s the first safe step toward feeling heard.

How we built it

MannSahay follows a modular, scalable architecture:

Frontend: React for a smooth, familiar user experience
Backend: FastAPI for fast and reliable APIs
Authentication & Storage:Firebase Auth + Firestore
AI Layer:
- Online: Gemini Flash via LangChain
- Offline: Ollama with Gemma models
Real-time Chat: Secure WebSockets
Cloud Infrastructure: Google Cloud Run, Functions, and Storage

Every activity link and prompt was curated under the guidance of a professional psychologist to ensure emotional safety.

Challenges we ran into

Balancing empathy and safety: Making the bots emotionally warm without crossing into unsafe advice. -Stigma-aware design: Ensuring the app feels non-clinical and non-judgmental. -Anonymity vs responsibility: Designing Incognito Mode while still maintaining crisis escalation paths.
Offline AI constraints: Optimizing local models to run smoothly on low-resource devices.
Tone switching: Making MannSahay feel like a calm guide and MannMitra feel like a real friend—without overlap.