Skip to content
BranchesTags
 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

14 Commits
backend
backend
 
 
frontend
frontend
 
 
.gitignore
.gitignore
 
 
README.md
README.md
 
 
vercel.json
vercel.json
 
 

Repository files navigation

RiseAgent AI

Rise.AI — Intelligent Multilingual Voice Agent

Every lead. Every language. Every time.

Python React FastAPI Tailwind License

🚀 What is RiseAgent?

RiseAgent is a production-grade, multilingual AI voice agent system built for Rupeezy's Authorized Person (AP) lead conversion program. It automatically calls every new lead within 5 minutes, conducts intelligent sales conversations in 7 Indian languages, handles objections with RAG-backed rebuttals, scores leads as Hot/Warm/Cold, and routes them to RM queues or WhatsApp follow-ups.

82% of sales leads go cold before anyone picks up the phone. RiseAgent calls every lead within 5 minutes — in their language, with the right answer, every time.

Built by Team MINDBOT at Technocrats Institute of Technology, Bhopal.

✨ Key Features

Feature Description
🗣️ Multilingual AI Calls Speaks Hindi, Tamil, Telugu, Marathi, Gujarati, Bengali & English — detects language automatically
⚡ 5-Minute Response Calls every lead within 5 minutes of arrival — no lead goes cold
🧠 Context-Aware Conversations RAG over Rupeezy's AP sales script; handles all 5 core objections
🔀 Code-Mix Detection Switches language mid-call if the lead code-mixes
📊 Weighted Lead Scoring 5-signal scoring engine: Hot → RM queue, Warm → WhatsApp, Cold → Re-engage
🤖 Dual LLM Failover Powered by Groq + Gemini with automatic failover for 99.9% uptime
📱 Real-Time Dashboard Live analytics, call transcripts, RM queue, and provider health monitoring
🔒 Demo Mode Full pipeline simulation without real calls — perfect for demos and testing

🏗️ Architecture

┌──────────────────────────────────────────────────────────────────┐
│                        FRONTEND (React 18)                       │
│  Landing Page ─── Dashboard ─── Leads ─── RM Queue ─── Call View │
└──────────────────────────┬───────────────────────────────────────┘
                           │ REST API
┌──────────────────────────▼───────────────────────────────────────┐
│                      BACKEND (FastAPI)                            │
│                                                                   │
│  POST /leads/new → Insert lead → Trigger call (background)       │
│                                       │                           │
│                     ┌─────────────────▼──────────────────┐       │
│                     │ [Demo] Simulate conversation        │       │
│                     │ [Prod] Exotel/Twilio outbound       │       │
│                     └─────────────────┬──────────────────┘       │
│                                       │                           │
│                     Voice loop: STT → LLM → TTS                  │
│                            (Sarvam AI)  (Groq/Gemini)             │
│                                       │                           │
│                     Objection detection + RAG rebuttal            │
│                            (ChromaDB + HuggingFace)               │
│                                       │                           │
│                     Score lead (Hot / Warm / Cold)                │
│                                       │                           │
│            Hot → RM Queue  │  Warm → WhatsApp  │  Cold → Re-engage│
└──────────────────────────────────────────────────────────────────┘

🛠️ Tech Stack

Layer Technology
Backend FastAPI, Uvicorn, Pydantic v2, Python 3.11+
LLM Groq (Llama 3) + Google Gemini with auto-failover via LangChain
RAG ChromaDB + HuggingFace multilingual embeddings
Voice Sarvam AI (STT + TTS + Translate)
Phone Exotel / Twilio (configurable)
WhatsApp Meta Cloud API (Graph API v18.0)
Database Supabase (PostgreSQL) or in-memory for demo
Frontend React 18, Vite 5, Tailwind CSS 3, Recharts, TanStack Query, Zustand
Animations Framer Motion, Lenis (smooth scroll)

📁 Project Structure

Rise-Agent/
├── backend/
│   ├── main.py                  # FastAPI app entry point
│   ├── config.py                # Environment & settings
│   ├── database.py              # Supabase / in-memory DB layer
│   ├── requirements.txt
│   ├── models/                  # Pydantic data models
│   ├── routers/                 # API route handlers
│   ├── services/
│   │   ├── call_service.py      # Outbound call orchestration
│   │   ├── conversation.py      # Multi-turn conversation engine
│   │   ├── language.py          # Language detection & switching
│   │   ├── llm.py               # LLM provider (Groq/Gemini failover)
│   │   ├── memory.py            # Conversation memory & context
│   │   ├── objection_handler.py # RAG-backed objection rebuttals
│   │   ├── scoring.py           # 5-signal lead scoring engine
│   │   ├── sarvam.py            # Sarvam AI voice integration
│   │   ├── rm_handoff.py        # RM queue routing
│   │   └── whatsapp.py          # WhatsApp follow-up service
│   ├── knowledge/               # RAG knowledge base documents
│   ├── prompts/                 # LLM system prompts
│   └── demo/                    # Demo personas & simulation
│
├── frontend/
│   ├── src/
│   │   ├── pages/
│   │   │   ├── LandingPage.jsx  # Cinematic landing experience
│   │   │   ├── Dashboard.jsx    # Real-time analytics dashboard
│   │   │   ├── Leads.jsx        # Lead management table
│   │   │   ├── RMQueue.jsx      # RM handoff queue
│   │   │   └── CallDetail.jsx   # Individual call transcript view
│   │   ├── components/
│   │   │   └── landing/         # Landing page components
│   │   │       ├── VideoBackground.jsx
│   │   │       ├── LiquidGlassCard.jsx
│   │   │       ├── WordsPullUp.jsx
│   │   │       ├── ScrollRevealText.jsx
│   │   │       ├── AboutSection.jsx
│   │   │       ├── FeaturesSection.jsx
│   │   │       ├── Navbar.jsx
│   │   │       └── LenisProvider.jsx
│   │   └── index.css
│   ├── tailwind.config.js
│   └── vite.config.js
│
└── README.md

⚡ Quick Start

Prerequisites

  • Python 3.11+
  • Node.js 18+
  • npm or yarn

1. Clone the repository

git clone https://github.com/Amit-Mahaseth/Rise-Agent.git
cd Rise-Agent

2. Backend Setup

cd backend
python3.11 -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt
cp .env.example .env
# Configure your API keys in .env
uvicorn main:app --reload

3. Frontend Setup

cd frontend
npm install
npm run dev

The app will be available at http://localhost:5173

4. Test Endpoints

# Health check
curl http://localhost:8000/health

# Seed all 5 demo personas
curl -X POST http://localhost:8000/leads/demo/seed

# Create a single lead (triggers demo call simulation)
curl -X POST http://localhost:8000/leads/new \
  -H "Content-Type: application/json" \
  -d '{"name": "Priya Sharma", "phone": "+919876543211", "source": "test"}'

# Dashboard stats
curl http://localhost:8000/dashboard/stats

# List all leads
curl http://localhost:8000/leads

# RM queue
curl http://localhost:8000/dashboard/rm-queue

🎭 Demo Mode

Set DEMO_MODE=true in .env (default). In demo mode:

  • ✅ No real phone calls are made
  • ✅ Full conversation pipeline runs with real LLM calls
  • ✅ Lead responses are simulated based on persona personality
  • ✅ Full scoring, memory, and routing work end-to-end
  • ✅ Dashboard updates in real-time

Demo Personas

Name Language Personality Expected Score
🇮🇳 Rajesh Kumar Hindi Skeptical Warm
🇬🇧 Priya Sharma English Enthusiastic Hot
🇮🇳 Karthik Rajan Tamil Neutral Warm
🇮🇳 Meena Patil Marathi Disengaged Cold
🇮🇳 Arjun Reddy Telugu Interested Hot

🔐 Environment Variables

See backend/.env.example for the full list. Minimum required for demo mode:

Variable Required Description
DEMO_MODE Yes Set to true for demo simulation
GROQ_API_KEY Yes Primary LLM provider
GOOGLE_API_KEY Yes Failover LLM provider (Gemini)
SARVAM_API_KEY Yes Voice services (STT/TTS/Translate)
SUPABASE_URL No PostgreSQL database (uses in-memory if unset)
SUPABASE_KEY No Supabase service role key

🎨 Landing Page

The landing page features a cinematic, studio-grade design with:

  • HLS video background with gradient overlays and grid lines
  • Liquid glass card with glassmorphism effects
  • Framer Motion animations — staggered word pull-ups, scroll-linked opacity reveals
  • Lenis smooth scrolling for buttery-smooth navigation
  • Responsive 12-column grid layout with mobile-first breakpoints

🤝 Team MINDBOT

Built at Technocrats Institute of Technology, Bhopal using:

  • Groq — Ultra-fast LLM inference
  • Google Gemini — Failover LLM with multimodal capabilities
  • Sarvam AI — Indian-language speech & translation
  • LangChain — LLM orchestration & RAG pipeline
  • React 18 — Modern, reactive frontend

📄 License

MIT License — see LICENSE for details.

About

https://rise-agent-one.vercel.app/

Resources

Readme
Activity

Stars

0 stars

Watchers

0 watching

Forks

0 forks
Report repository

Releases

Packages

Contributors

Languages