Bharathi 🌟

Build for India Hackathon | Organized by Kerala Startup Mission (KSUM)

Project by Team Loreal

---

👥 Team

Team Name: Loreal

Name	Role
Aibel Bin Zacariah	Team Leader
Aravind Manoj	Team Member
Abin Thomas	Team Member
Devadarsan C P	Team Member

---

📖 Project Description

🔍 The Problem

UPI made digital payments universally accessible by removing technical complexity. Any business small or large can integrate payments without building financial infrastructure from scratch.

Intelligent voice agents do not offer the same accessibility.

Deploying a voice agent today requires custom development, AI model integration, telephony setup, API orchestration, and infrastructure management. This demands AI engineers and significant financial investment.

As a result, most businesses — including large platforms — do not implement intelligent voice agents. The capability exists, but the deployment layer remains complex and inaccessible.

💡 The Solution

Bharathi is the UPI moment for voice AI.. It is a no-code platform that enables any business — from a street-scale vendor to a large enterprise — to deploy a fully functional AI voice agent in just 2–3 clicks or trigger one directly through a call. No code, no AI engineers, and no infrastructure overhead. Bharathi removes barriers of:

🌐 Language Multi language support

🧓 Age & Tech Literacy Voice first, no digital navigation required

📱 Platform No app downloads

🏪 Business Size Scales from local shops to enterprises

At its core, Bharathi runs on an agentic AI architecture where autonomous agents understand context, invoke tools, maintain memory, and execute multi-step workflows — handling bookings, queries, calls, and task coordination end-to-end.

🤖 Agentic AI Capabilities

• Autonomous Orchestration — Agents independently plan and execute multi-step tasks
• Tool Use — Agents call external tools and APIs (voice, image gen, search) to complete goals
• Persistent Memory — Supermemory gives agents long-term context across sessions
• Multi-modal — Agents work across text, voice, and images
• Observability — Full tracing of agent reasoning and actions via Langfuse + OpenTelemetry

⚙️ Platform Integrations

• 🤖 AI Models — Google Gemini & Groq (Llama) for language understanding and generation
• 🧠 Memory Layer — Supermemory for persistent AI context and tool access
• 📞 Voice Calls — Twilio integration for call-based agent interactions
• 🖼️ Image Generation — When a business adds a new offer or seasonal promotion, Bharathi automatically generates a professional poster and posts it to the company's social media — no designer, no manual effort required
• 📊 Dashboard — A rich, interactive dashboard with real-time data
• 🔐 Authentication — Secure auth with better-auth
• ⚡ Background Jobs — Inngest for reliable background processing
• 🗄️ Database — PostgreSQL with Drizzle ORM
• ☁️ Storage — AWS S3 for file/media storage
• 📡 API Layer — tRPC for end-to-end type-safe APIs

---

🛠️ Tech Stack

Frontend

• Next.js 16 (React 19)
• Tailwind CSS v4
• shadcn/ui + Radix UI
• Framer Motion for animations
• tRPC + TanStack Query

Backend

• Python (FastAPI / scripts)
• Drizzle ORM + PostgreSQL
• Inngest for background jobs
• Better Auth for authentication

AI / Integrations

• Vercel AI SDK
• Google Gemini (@ai-sdk/google)
• Groq / Llama (@ai-sdk/groq)
• Supermemory
• Sarvam AI (sarvamai)
• Twilio for voice
• AWS S3 for storage
• Upstash Redis

---

🚀 Getting Started

Prerequisites

• Node.js >= 18 and Bun (or npm/yarn/pnpm)
• Python >= 3.11 and uv (Python package manager)
• PostgreSQL database
• Required API keys (see Environment Variables below)

Installation

1. Clone the repository:

   git clone <repo-url>
   cd bharathi

2. Install Node dependencies:

   bun install
   # or
   npm install

3. Install Python dependencies:

   uv sync

4. Set up environment variables (see Environment Variables)

5. Run database migrations:

   bun run db:migrate
   # or
   npx drizzle-kit migrate

Running the Development Server

bun dev
# or
npm run dev

Open http://localhost:3000 in your browser.

Running the Python Backend

cd backend
uv run python main.py

---

🔑 Environment Variables

Create a .env.local file in the root directory with the following:

# Database
DATABASE_URL=

# Auth
BETTER_AUTH_SECRET=

# AI
GOOGLE_GENERATIVE_AI_API_KEY=
GROQ_API_KEY=

# Twilio
TWILIO_ACCOUNT_SID=
TWILIO_AUTH_TOKEN=
TWILIO_PHONE_NUMBER=

# AWS S3
AWS_ACCESS_KEY_ID=
AWS_SECRET_ACCESS_KEY=
AWS_REGION=
AWS_BUCKET_NAME=

# Upstash Redis
UPSTASH_REDIS_REST_URL=
UPSTASH_REDIS_REST_TOKEN=

# Supermemory
SUPERMEMORY_API_KEY=

# Langfuse (Observability)
LANGFUSE_PUBLIC_KEY=
LANGFUSE_SECRET_KEY=
LANGFUSE_HOST=

# Sarvam AI
SARVAM_API_KEY=

---

📁 Project Structure

bharathi/
├── app/                  # Next.js App Router pages
│   ├── api/              # API routes
│   ├── dashboard/        # Dashboard pages
│   ├── call/             # Call-related pages
│   ├── login/            # Auth pages
│   └── register/
├── backend/              # Python backend scripts
│   ├── main.py
│   ├── contextGen.py
│   └── imageGen.py
├── components/           # Shared React components
├── db/                   # Database schema & config
├── drizzle/              # Drizzle migrations
├── inngest/              # Background job functions
├── lib/                  # Shared utilities
├── platform/             # Platform-specific logic
├── trpc/                 # tRPC router definitions
└── public/               # Static assets

---

📜 Available Scripts

Command	Description
bun dev	Start development server
bun build	Build for production
bun start	Start production server
bun lint	Run ESLint

---