✨ Features

• 🚀 One-line install — import "autoverse-agent" and the widget appears
• 🧠 Auto-learns your site — Scrapes products, pages, and APIs automatically
• 🛒 Order automation — Collects customer details and places orders step-by-step
• 🎨 Fully customizable — Colors, theme, position, tone via admin dashboard
• 🌍 Bilingual — English + Urdu with auto-detection
• 🔑 Multi-tenant — One server, unlimited chatbots with unique Bot IDs
• 📡 Network interception — Auto-indexes API responses for richer context
• 💬 Human-like tone — Sounds like a luxury store concierge, not a robot

🚀 Quick Start

1. Install the package

npm install autoverse-agent

2. Add one line to your app

import "autoverse-agent";

That's it. A chat bubble appears on your website. Click it, enter your Bot ID, and your AI assistant is live.

📋 How It Works

┌──────────────────────────────────────────────────┐
│            AUTOVERSE SERVER (deploy once)         │
│                                                   │
│   Admin Dashboard  ◄────►  API + AI Engine        │
│   (create & manage bots)   (Groq LLM + RAG)      │
└──────────────────┬───────────────────────────────┘
                   │
      ┌────────────┼────────────┐
      ▼            ▼            ▼
  Website A    Website B    Website C
  botId: abc   botId: def   botId: ghi
  💬 Widget    💬 Widget    💬 Widget

1. Deploy the Autoverse server once (locally or on Railway/Render/VPS)
2. Create a bot in the admin dashboard — get a unique Bot ID
3. Install the npm package on any website
4. Enter the Bot ID in the widget — your AI assistant starts working

🔧 Integration Methods

Method 1: NPM Package (Recommended)

npm install autoverse-agent

Zero-config (setup via widget):

import "autoverse-agent";

Pre-configured (skip setup screen):

import { initWidget } from "autoverse-agent";

initWidget({ botId: "YOUR_BOT_ID" });

Method 2: Script Tag (No npm needed)

<script src="https://YOUR-SERVER.com/autoverse/widget.js?botId=YOUR_BOT_ID"></script>

Method 3: Meta Tag + Import

<meta name="autoverse-bot-id" content="YOUR_BOT_ID">

import "autoverse-agent";

Method 4: Express Middleware

import express from "express";
import { autoverse } from "autoverse-agent";

const app = express();
app.use(autoverse());
app.listen(3000);

🎛️ Admin Dashboard

The Autoverse dashboard lets you manage multiple chatbots from one place.

Features

| Tab | What You Can Do | |-----|----------------| | 🎨 Appearance | Primary/accent colors, dark/light theme, position, bubble size, welcome message | | 🗣️ Tone | Language preference, response length, custom system prompt | | 🔗 Integration | Copy-paste embed codes, view Bot ID, creation date |

Live Preview

Changes are applied instantly — no redeployment needed. The widget fetches its configuration from the server on every load.

🛒 Order Automation

The chatbot handles the full order flow automatically:

Customer: "I want to buy the Swiss Army Knife"

Bot: "Great choice! 🎉 The Swiss Army Knife is $49.99.
      How many would you like?"

Customer: "2 please"

Bot: "Perfect! Could I get your full name for the order?"

Customer: "Ahmed Khan"

Bot: "Thanks Ahmed! And your phone/WhatsApp number?"

Customer: "0300-1234567"

Bot: "What's the delivery address?"

Customer: "House 12, Street 5, DHA Phase 2, Lahore"

Bot: "Would you like Cash on Delivery or online payment?"

Customer: "COD"

Bot: "Let me confirm your order:
      📦 Product: Swiss Army Knife x 2
      💸 Total: $99.98
      👤 Name: Ahmed Khan
      📱 Phone: 0300-1234567
      📍 Address: House 12, Street 5, DHA Phase 2, Lahore
      💳 Payment: Cash on Delivery

      Shall I place this order for you?"

Customer: "Yes"

Bot: "✅ Order Placed Successfully!
      🎉 Order #AV-482916
      Your order will be processed shortly.
      By the way, you might also love these..."

🖥️ Server Setup

Running Locally

# Clone the repo
git clone https://github.com/your-repo/autoverse-agent.git
cd autoverse-agent

# Install dependencies
cd autoverse-agent && npm install
cd ../dashboard && npm install

# Build everything
cd ../autoverse-agent && npm run build
cd ../dashboard && npx vite build

# Start the server
cd ../autoverse-agent && npm run serve

Dashboard opens at http://localhost:3000

Environment Variables

Create a .env file in the autoverse-agent directory:

# Required
GROQ_API_KEY=gsk_your_api_key_here

# Optional
GROQ_MODEL=llama3-8b-8192      # AI model (default: llama3-8b-8192)
PORT=3000                       # Server port (default: 3000)
PUBLIC_URL=http://localhost:3000 # Your server's public URL

| Variable | Default | Description | |----------|---------|-------------| | GROQ_API_KEY | — | Your Groq API key (get one free) | | GROQ_MODEL | llama3-8b-8192 | LLM model to use | | PORT | 3000 | Server port | | PUBLIC_URL | http://localhost:3000 | Public URL (set this when deploying) |

Supported Models

| Model | Speed | Quality | Best For | |-------|-------|---------|----------| | llama3-8b-8192 | ⚡ Fast | Good | Most use cases | | llama3-70b-8192 | Medium | Excellent | Complex queries | | mixtral-8x7b-32768 | Medium | Great | Long conversations | | gemma-7b-it | ⚡ Fast | Good | Simple Q&A |

🚢 Deploy to Production

Railway

1. Push the repo to GitHub
2. Connect to Railway
3. Set environment variables:

   GROQ_API_KEY=gsk_...
   PUBLIC_URL=https://your-app.railway.app

4. Set build command:

   cd autoverse-agent && npm install && npm run build && cd ../dashboard && npm install && npx vite build

5. Set start command:

   cd autoverse-agent && node -e "require('dotenv').config(); require('./dist/server').startServer()"

Render / Fly.io / VPS

Same pattern — set PUBLIC_URL to your deployed URL, build both projects, and start the server.

Important: After deploying, update the AUTOVERSE_CLOUD_URL constant in widget.ts to your production URL, rebuild, and republish the npm package. This makes import "autoverse-agent" connect to your server automatically.

📦 API Reference

initWidget(options?)

Initialize the chat widget manually.

import { initWidget } from "autoverse-agent";

initWidget({
  botId: "abc123",              // Optional: skip setup screen
  serverUrl: "https://my.server" // Optional: custom server URL
});

| Option | Type | Default | Description | |--------|------|---------|-------------| | botId | string | "" | Bot ID from dashboard. If empty, shows setup screen | | serverUrl | string | Built-in cloud URL | Autoverse server URL |

autoverse(config?)

Express middleware for server-side integration.

import { autoverse } from "autoverse-agent";

app.use(autoverse({
  groqModel: "llama3-8b-8192",
  groqBaseUrl: "https://api.groq.com/openai/v1",
  verbose: true
}));

startServer(config?)

Start a standalone Autoverse server with built-in dashboard.

import { startServer } from "autoverse-agent";

startServer({
  port: 3000,
  publicUrl: "https://my-server.com",
  groqModel: "llama3-70b-8192"
});

🔌 REST API Endpoints

| Method | Endpoint | Description | |--------|----------|-------------| | GET | /autoverse/widget.js | Widget JavaScript (add ?botId=xxx) | | GET | /autoverse/api/health | Server health check | | POST | /autoverse/api/init | Initialize agent for a website | | POST | /autoverse/api/chat | Send a chat message | | GET | /autoverse/api/bots/:id/config | Get bot appearance config | | POST | /autoverse/api/dashboard/bots | Create a new bot | | GET | /autoverse/api/dashboard/bots | List all bots | | GET | /autoverse/api/dashboard/bots/:id | Get bot details | | PUT | /autoverse/api/dashboard/bots/:id | Update bot settings | | DELETE | /autoverse/api/dashboard/bots/:id | Delete a bot |

🤝 How the NPM Package Connects to the Dashboard

  Developer's Website                    Your Autoverse Server
  ┌─────────────────┐                   ┌─────────────────────┐
  │                 │                   │                     │
  │  import         │   Bot ID          │   Admin Dashboard   │
  │  "autoverse-    │ ◄──────────────── │   (create bots,     │
  │   agent"        │                   │    customize look)   │
  │                 │                   │                     │
  │  💬 Widget      │ ──── API ────►    │   AI Engine         │
  │  appears!       │ ◄── Response ──── │   (Groq + RAG)      │
  │                 │                   │                     │
  └─────────────────┘                   └─────────────────────┘

1. Admin creates a bot in the dashboard → gets a Bot ID
2. Developer installs autoverse-agent on their website
3. Widget auto-appears → user enters the Bot ID
4. Widget connects to the server → AI starts helping customers
5. Admin can change appearance/tone anytime from the dashboard — changes apply instantly

📝 License

MIT © Autoverse