Desi AI Chatbot – Complete Technical Documentation

Desi AI Chatbot

Complete Technical Architecture, AI Flow & Development Documentation

chatwithvanshrajput.in ek fully-functional, production-grade AI chatbot hai jo maine (Vansh) ne scratch se design, develop aur deploy kiya hai. Yeh sirf ek chatbot nahi, balki ek full-stack AI system hai jo real users ke liye bana hai.

🤖 Open Live AI Chatbot
Full-Stack AI-Powered Multi-Modal Multi-Language Secure & Scalable
0

Lines of Code

0

Edge Functions

0

Database Tables

0

AI Models

Overview
Features
Architecture
AI & ML
Security
Development Journey

Desi AI Chatbot ek multi-modal conversational system hai. User text, voice ya image bhej sakta hai, aur chatbot us input ko samajhkar Hinglish mein natural response deta hai.

System ko is tarah design kiya gaya hai ki har request secure ho, scalable ho, aur future mein naye AI models easily add kiye ja sakein.

Hinglish Conversations

Hindi + English mix, bilkul real desi tone mein.

Image Understanding

User images ko analyze karke meaningful answers.

Real-Time Web Search

Latest information using Perplexity API.

Image Generation

Text prompts se AI images generate karna.

Voice Interaction

Speech-to-Text & Text-to-Speech (ElevenLabs).

Role-Based Access

Free, Mentorship & Admin access levels.

Gamified UI

Cyberpunk-style animations & effects.

Desi Personality

Friendly, fun & culturally relevant responses.

System Architecture Flow


User Input
Text / Voice / Image
React Frontend
Validation + UI Logic
Edge Functions
Secure API Gateway
AI Layer
Gemini / Perplexity
Response Engine
Hinglish Formatting

AI layer is the heart of this system. Primary model Google Gemini hai jo reasoning, conversation aur image understanding handle karta hai.

Perplexity API ko web-search aur fallback model ke roop mein use kiya gaya hai, taaki chatbot hamesha up-to-date information de sake.

Voice ke liye ElevenLabs integrate kiya gaya hai jisse natural human-like speech possible hoti hai.

Security system ka sabse important part hai. Isliye maine Supabase Authentication, Row Level Security (RLS), aur Edge Functions use kiye hain.

AI keys frontend mein kabhi expose nahi hoti. Har request pehle Edge Function se pass hoti hai, jahan validation aur permission checks hote hain.

Phase 1

Planning

Personality, language & UX define ki.

Phase 2

Frontend

React + TypeScript UI build ki.

Phase 3

Backend

Supabase + Edge Functions.

Phase 4

AI

Gemini, Perplexity & ElevenLabs.

Phase 5

Polish

Testing, performance & gamification.

Built & Architected by Vansh Raj Singh
chatwithvanshrajput.in
Desi AI Chatbot – Backend Documentation

🔒 Backend Documentation

Enter 4-digit PIN to unlock