@_401_nick/lm-vision

VisionFramework

VisionFramework is a Node.js library that leverages the OpenAI API, specifically the GPT-4-Vision model, to analyze images and generate useful text-based outputs such as descriptions, tags, captions, alt text, and stories from images. It provides an easy-to-use interface for integrating advanced image analysis and text generation capabilities into your projects.

Features

• Image Descriptions: Generate detailed descriptions of images.
• Image Tags: Create a list of tags relevant to an image.
• Query Answers: Answer specific queries about an image.
• Image Captions: Produce concise captions for images.
• Alt Text Generation: Generate alternative text for images for accessibility.
• Story Creation: Create narrative stories from a series of images.

Prerequisites

Before you begin, ensure you have met the following requirements:

• Node.js installed on your machine.
• An OpenAI API key with access to the GPT-4-Vision model.

Installation

To use VisionFramework in your project, follow these steps:

• Clone this repository or download the source code.
• Navigate to the project directory and install the dependencies:

  npm install @_401_nick/lm-vision
  

• Create a .env file in the project root and add your OpenAI API key:

  OPENAI_API_KEY=your_openai_api_key_here
  

Usage

Here's how to use VisionFramework in your project:

• Import the VisionFramework class from the module.

  const VisionFramework = require('./VisionFramework');
  

• Create an instance of the VisionFramework.

  const vision = new VisionFramework();
  

• Call the methods provided by the VisionFramework instance to analyze images and generate text. Here are some examples:

  // Generate an image description
  vision.analyzeImageDescription('https://example.com/image.jpg')
      .then(description => console.log(description))
      .catch(error => console.error(error));
  
  // Generate tags for an image
  vision.generateImageTags('https://example.com/image.jpg')
      .then(tags => console.log(tags))
      .catch(error => console.error(error));