BuildbyManoj Chatbot

Link: https://buildbymanoj-chatbot.onrender.com

A modern, responsive web-based chatbot with session management and RAG (Retrieval-Augmented Generation) capabilities. This Flask-powered AI assistant uses OpenRouter's DeepSeek-V3 model and scrapes DuckDuckGo for real-time information.

Features

• Session-Based Memory: The chatbot remembers your entire conversation during a browser session
• RAG (Retrieval-Augmented Generation): Automatically searches the web for real-time information when needed
• Responsive UI: Clean, modern interface that works on both desktop and mobile devices
• Markdown Support: Bot responses support markdown formatting for better readability
• Smart Web Search: Intelligently decides when to use web search based on the query

Technologies Used

• Backend: Flask (Python)
• Frontend: HTML, CSS, JavaScript
• AI: DeepSeek Chat model via OpenRouter API
• Web Scraping: BeautifulSoup4 for DuckDuckGo search results
• Session Management: Flask-Session for browser session-based conversation history

Getting Started

Prerequisites

• Python 3.7+
• OpenRouter API key

Installation

1. Clone this repository:

   git clone https://github.com/buildbymanoj/BuildbyManoj-chatbot.git
   cd BuildbyManoj-chatbot
   

2. Install dependencies:

   pip install -r requirements.txt
   

3. Create a .env file in the project root with your API key:

   OPENROUTER_API_KEY=your_api_key_here
   

4. Run the application:

   python app1.py
   

5. Open your browser and navigate to http://localhost:10000

How It Works

1. User Input: User sends a message through the web interface
2. RAG Decision: System decides whether web search is needed based on keywords and query complexity
3. Web Search: If needed, retrieves relevant information from DuckDuckGo
4. API Call: Sends user message, web search results (if any), and conversation history to the AI
5. Response: AI generates a response which is displayed to the user
6. Session Storage: Both the user message and AI response are stored in the session for context

Deployment

The app is already deployed on Render.com and includes a render.yaml file for easy deployment.

License

This project is open-source and available for personal and commercial use.

Credits

Developed by BuildbyManoj