Vedanta

Vedanta is a web-based chat interface for interacting with local AI models running on your device. It provides a sleek, responsive UI with support for Markdown formatting, code block syntax highlighting, message history, and streaming responses—making it ideal for experimenting with and using local LLMs (Large Language Models).

Features

• Chat Interface: Modern chat UI for conversational interaction.
• Local Model Selection: Choose different local AI models, or let the app auto-select.
• Markdown & Code Support: Messages are rendered with Markdown, including code blocks and syntax highlighting.
• Streaming Responses: Simulated streaming output for a more interactive chat feel.
• Copy Code Button: Easily copy code from AI responses.
• Welcome Examples: Quick example prompts to get started (e.g., Python code generation, math help, concept explanations).
• History Panel: Toggleable sidebar for browsing previous conversations.

Getting Started

1. Clone the repository

   git clone https://github.com/suhanpahari/Vedanta.git
   cd Vedanta

2. Run the Application

   • Vedanta uses a Flask backend to serve the chat API.
   • Make sure you have Python installed.
   • Install dependencies:

     pip install flask

   • Start the server:

     flask run

   • Open in.html or templates/index.html in your browser.

3. Configure Local Models

   • Select your preferred local AI model from the dropdown in the interface.
   • The app will use the selected model for your queries.

Project Structure

• in.html, templates/index.html: Main HTML files for the chat interface.
• static/script.js: Core frontend logic (event handling, message streaming, Markdown parsing, etc.).
• static/: Static assets (JS, CSS).
• api/chat: Flask backend endpoint for processing chat messages.

Usage

• Type your message in the input box and press "Send" or Enter.
• Explore example prompts for inspiration.
• Select different models to compare outputs.
• Code snippets in responses can be copied with the button provided.

Technologies Used

• Frontend: HTML, CSS, JavaScript (vanilla)
• Backend: Python (Flask)
• Markdown Parsing: Custom logic (can be replaced with libraries like marked.js)
• Syntax Highlighting: Prism.js (integration suggested)
• Clipboard: Uses browser Clipboard API

Customization

• Add new models or integrate with different local LLM backends.
• Enhance Markdown parsing or switch to a more robust library.
• Customize UI themes and layouts in in.html and CSS.

License

This project currently does not specify a license. Please add one if you intend to share or modify the code.

Author

Developed by @suhanpahari.

---

Note: This project is best suited for experimentation with local AI models. Online model support or advanced features may require additional setup.