Skip to content
/ ai-coder Public

A website built with Next.js that integrates Meta's LLM model, Llama 3.1 405B, to generate React.js components using prompts.

Notifications You must be signed in to change notification settings

jei3m/ai-coder

Repository files navigation

AI Coder

AI Coder is a simple website created using Next.js, integrated with Llama 405B for React.js code generation. This project was developed as an experiment to explore the capabilities of Llama 405B. This is heavily inspired by LlamaCoder. However, instead of using Together.AI, I opted for the SambaNova for cloud-based inference of Llama 405B.

Features

  • Next JS: Utilizes a front-end library Next JS for building user interfaces.
  • Llama 405B AI Integration: Generates code in React JS with the use prompts.
  • SambaNova: the fastest, cloud-based inference platform on Llama 405B.

Purpose

The main purpose of AI-Coder is to explore and understand how Meta's Llama 405B can be integrated into web applications for coding tasks. Built with Next.js, this project aims to study the practical applications and limitations of AI-driven code generation, focusing on creating React.js components from prompts.

Installation

To get started with the AI Coder, follow these steps:

  1. Clone the repository:

    git clone https://github.com/jei3m/ai-coder.git
  2. Navigate to the project directory:

    cd ai-coder
  3. Install the dependencies:

    npm install
  4. Set Up Environment Variables Create a .env file in the root directory and add your Sambanova API credentials.

    SAMBANOVA_API_KEY=your_sambanova_api_key
  5. Start the development server:

    npm run dev
    

Usage

Once the development server is running, open your browser and navigate to http://localhost:3000. You can start typing prompts and the integrated AI will automatically create React JS components for you.

About

A website built with Next.js that integrates Meta's LLM model, Llama 3.1 405B, to generate React.js components using prompts.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published