🚀 OneClickAgent

WhatsApp-first AI business automation for Indian small businesses. 100% FREE to run.

Customer WhatsApp pe message karta hai → AI Hindi mein reply karti hai → Appointment book hoti hai → Payment collect hota hai. Sab autopilot pe.

💰 Total Running Cost: ₹0

Service	Provider	Cost
AI Brain	Groq (LLaMA 3.1 70B)	FREE
AI Fallback	DeepSeek	~₹0.07/query
Voice STT	Groq Whisper	FREE
Voice TTS	Google Cloud TTS	FREE (1M chars/month)
Database	Neon.tech PostgreSQL	FREE (512MB)
Knowledge Base	In-memory + disk	FREE
Hosting	Railway.app / Render	FREE tier
WhatsApp	Meta Cloud API	FREE (1000 chats/month)

⚡ Deploy in 5 Minutes

Step 1: Get FREE API Keys (5 min)

Service	Sign Up	What You Get
Groq	console.groq.com	Free API key → GROQ_API_KEY
Neon	neon.tech	Free PostgreSQL → DATABASE_URL
Meta	developers.facebook.com	WhatsApp API → WHATSAPP_TOKEN + PHONE_NUMBER_ID

Step 2: Deploy on Railway (FREE)

1. Push code to GitHub
2. Go to railway.app → New Project → Deploy from GitHub
3. Add environment variables (from .env.example)
4. Done! Get your URL: https://your-app.railway.app

Step 3: Connect WhatsApp

1. Meta Developer Dashboard → WhatsApp → Configuration
2. Webhook URL: https://your-app.railway.app/webhook
3. Verify Token: same as WHATSAPP_VERIFY_TOKEN in env
4. Subscribe to: messages

OR: Deploy on Render (FREE)

1. Push to GitHub
2. render.com → New Web Service → Connect repo
3. It auto-detects render.yaml — just add env vars
4. Done!

OR: Run Locally

git clone <your-repo>
cd OneClickAgent
cp .env.example .env    # Fill in your keys
npm install
npx prisma generate
npx prisma migrate dev
npm run dev

🏗️ Architecture

Customer WhatsApp Message
        ↓
   Webhook (Express)
        ↓
   Agent Router (intent detection)
        ↓
   ┌─────────────┬──────────┬──────────────┐
   │ Receptionist │  Sales   │   Support    │
   │   Agent      │  Agent   │   Agent      │
   └──────┬───────┴────┬─────┴──────┬───────┘
          ↓            ↓            ↓
   AI Engine (Groq FREE → DeepSeek fallback)
          ↓
   Knowledge Base (in-memory, persisted to disk)
          ↓
   WhatsApp Reply (text + voice + buttons)

🔌 API Endpoints

Auth

Method	Endpoint	Description
POST	/api/auth/register	Register business
POST	/api/auth/login	Login → JWT token
GET	/api/auth/profile	Get profile

Dashboard (JWT required)

Method	Endpoint	Description
GET	/api/dashboard/stats	Analytics overview
GET	/api/dashboard/conversations	List conversations
GET	/api/dashboard/conversations/:id	Conversation + messages
GET	/api/dashboard/customers	Customer list + search
GET	/api/dashboard/bookings	All bookings
GET	/api/dashboard/payments	All payments
GET	/api/dashboard/engine	AI engine health

Operations (JWT required)

Method	Endpoint	Description
POST	/api/payments/create-link	Generate payment link
PATCH	/api/bookings/:id/cancel	Cancel booking
PATCH	/api/bookings/:id/reschedule	Reschedule booking
POST	/api/broadcast/send	Bulk WhatsApp broadcast
POST	/api/broadcast/follow-ups	Auto follow-ups

Knowledge Base (JWT required)

Method	Endpoint	Description
POST	/api/knowledge/upload	Upload PDF/Excel/Word
POST	/api/knowledge/text	Add text knowledge
GET	/api/knowledge/search?q=...	Search knowledge
DELETE	/api/knowledge	Clear all

🤖 Multi-Agent System

Agent	Handles	Trigger Keywords
Receptionist	Greet, route, book appointments	hi, namaste, book, appointment
Sales	Products, pricing, payment links	price, kitna, pay, Rs
Support	Complaints, escalation	problem, dikkat, refund, kharab

Auto-detects intent in Hindi, Hinglish, English, Tamil, Telugu. Human handoff triggers when AI confidence < 30%.

🗣️ Voice (All FREE)

• Groq Whisper → Customer voice note → text (FREE tier)
• Google TTS → AI reply → voice message (FREE 1M chars/month)
• gTTS fallback → Unlimited free TTS (Python)
• Auto voice reply when customer sends voice note

📁 Project Structure

OneClickAgent/
├── src/
│   ├── index.js                # Express server
│   ├── config/index.js         # Configuration
│   ├── engine/EngineService.js # AI brain (Groq/DeepSeek)
│   ├── agents/
│   │   ├── AgentRouter.js      # Intent → Agent
│   │   ├── ReceptionistAgent.js
│   │   ├── SalesAgent.js
│   │   └── SupportAgent.js
│   ├── services/
│   │   ├── WhatsAppService.js  # Meta Cloud API
│   │   ├── VoiceService.js     # Whisper + TTS
│   │   ├── BookingService.js   # Google Calendar
│   │   ├── PaymentService.js   # Razorpay + Stripe
│   │   ├── KnowledgeService.js # Vector search
│   │   └── BroadcastService.js # Bulk + cron
│   ├── routes/                 # 7 route files
│   ├── middleware/auth.js      # JWT auth
│   └── utils/logger.js
├── prisma/schema.prisma        # 6 DB tables
├── railway.json                # Railway deploy config
├── render.yaml                 # Render deploy config
├── docker-compose.yml          # Local Docker setup
├── Dockerfile
└── .env.example

🧪 Test Results: 33/33 Passed

Engine: 10/10 ✅  |  Agents: 10/10 ✅  |  WhatsApp: 4/4 ✅
Knowledge: 4/4 ✅  |  Auth: 4/4 ✅  |  API: 1/1 ✅

License

MIT