Introduction

Welcome to the @logai/openai library! This library is an enhanced client for the OpenAI API, equipped with powerful logging, tracking, and safety features provided by LogAI. With @logai/openai, you can seamlessly integrate OpenAI’s capabilities into your application while gaining additional functionalities to monitor, analyze, and block potentially abusive or harmful prompts.

Features

• Seamless Integration with OpenAI: Enjoy the full power of OpenAI's language models while adding essential observability tools.
• Track and Log Prompts: Automatically log every request and response for easy tracking and analysis.
• Abuse Detection: Protect your application from prompt injections and other harmful input with built-in pre-checks.
• Customizable Blocking: Automatically block abusive prompts and replace them with custom responses.
• Real-Time Analysis: Monitor prompt interactions in real-time to ensure compliance and safety.

Installation

To get started, install the @logai/openai package using npm or yarn:

npm install @logai/openai openai
# or
yarn add @logai/openai openai
# or
pnpm add @logai/openai openai

Getting Started

Basic Usage

import LogAI from "@logai/openai";

const logaiClient = LogAI({
  apiKey: "the-logai-api-key-for-your-organization",
  project: "your-logai-project-id",
  env: "dev", // or "prod" for production
  openai: {
    apiKey: OPENAI_API_KEY,
    // other OpenAI client options can be passed here
  },
});

async function getAIResponse(prompt: string) {
  try {
    const response = await logaiClient.chat.completions.create({
      model: "gpt-4o-mini",
      messages: [{ role: "user", content: prompt }],
    });

    console.log(response.data.choices[0].message.content);
  } catch (error) {
    console.error("Error with OpenAI:", error);
  }
}

getAIResponse("What's the weather like today?");

Advanced Usage

Tracking and Logging

@logai/openai automatically logs every request and response, storing essential details for later analysis. This is invaluable for debugging, auditing, or understanding user interactions with the model.

Abuse Detection and Blocking

The library includes built-in mechanisms to detect and block abusive or malicious prompts. You can customize the block messages or configure different thresholds for blocking based on your needs.

To customize the blocking behavior, you can update the Blocking Strategy in your organization's LogAI dashboard. Also, you can change the default block message to various responses.

Environment Variables

Ensure you have the following environment variables set:

• OPENAI_API_KEY: Your OpenAI API key.
• LOGAI_API_KEY: Your LogAI API key for enhanced logging and abuse detection.
• LOGAI_PROJECT_ID: Your LogAI project identifier.
• LOGAI_ENV: The environment in which the library is running (development, staging, production).

Contributing

We welcome contributions! If you'd like to contribute to @logai/openai, please follow these steps:

1. Fork the repository.
2. Create a new branch with your feature or fix: git checkout -b feature-branch.
3. Commit your changes: git commit -m "Add new feature".
4. Push to the branch: git push origin feature-branch.
5. Open a pull request.

License

This project is licensed under the Apache License. Please see the LICENSE file for more details.

By integrating @logai/openai into your application, you gain robust observability and safety features on top of OpenAI’s powerful language models, ensuring that your application remains secure and compliant while delivering exceptional AI-driven experiences.