Package Exports

agents-js
agents-js/src/index.js

This package does not declare an exports field, so the exports above have been automatically detected and optimized by JSPM instead. If any package subpath is missing, it is recommended to post an issue to the original package (agents-js) to support the "exports" field. If that is not possible, create a JSPM override to customize the exports field for this package.

Readme

agents-js

agents-js is a JavaScript package designed to facilitate interaction with chat models, such as OpenAI's GPT models. It provides a structured way to define agents with specific instructions and functions, and it handles tool calls and context variables efficiently.

Installation

To install agents-js, you can use npm:

npm install agents-js

Usage

The primary classes provided by the agents-js package are Agent, AgentController, Result, and Response. Here, we'll walk through some basic usage examples.

Example 1: Creating an Agent

const { Agent } = require("agents-js");

// Define a simple agent with a name and instruction
const myAgent = new Agent({
  name: "MyAssistant",
  instructions: "You are a friendly assistant.",
  model: "gpt-4",
});

console.log(myAgent);

Example 2: Initializing OpenAI Client and Using AgentController

Before using the AgentController class to interact with chat models, you need to initialize the OpenAI client with your API key.

const OpenAI = require("openai");
const { AgentController, Agent } = require("agents-js");

// Initialize OpenAI client
const client = new OpenAI({
  apiKey: "sk-thekey", // Replace 'sk-thekey' with your actual API key
});

// Use the initialized client to create a AgentController instance
const agentController = new AgentController(client);

// Create an agent
const agent = new Agent({
  name: "Assistant",
  instructions: "You are a helpful assistant.",
  model: "gpt-4",
});

// Define a message history
const messages = [{ role: "user", content: "Hello, can you help me?" }];

// Fetch a chat completion
(async () => {
  const response = await agentController.run(agent, messages);
  console.log(response);
})();

Example 3: Handling Tool Calls

You can define functions that the agent can call during its operation.

const { AgentController, Agent, Result } = require("agents-js");

// Define a simple function that the agent can use
function add({ a, b }) {
  return new Result({ value: `${a + b}` });
}

// Initialize AgentController
const agentController = new AgentController();

// Create an agent with the function
const agent = new Agent({
  name: "Calculator",
  instructions: "You can perform addition using the add function.",
  model: "gpt-4",
  functions: [add],
});

// Define a message history
const messages = [{ role: "user", content: "What is 2 + 3?" }];

// Fetch a chat completion
(async () => {
  const response = await agentController.run(agent, messages);
  console.log(response);
})();

Example 4: Using Context Variables

Context variables allow the agent to maintain state across interactions.

const { AgentController, Agent, Result } = require("agents-js");

// Define a function using context variables
function greet({ name }) {
  return new Result({ value: `Hello, ${name}!` });
}

// Initialize AgentController
const agentController = new AgentController();

// Create an agent with the function
const agent = new Agent({
  name: "Greeter",
  instructions: "You can greet people using the greet function.",
  model: "gpt-4",
  functions: [greet],
});

// Define a message history and context variables
const messages = [{ role: "user", content: "Greet John for me." }];
const contextVariables = { name: "John" };

// Fetch a chat completion
(async () => {
  const response = await agentController.run(agent, messages, contextVariables);
  console.log(response);
})();

Conclusion

agents-js provides a flexible framework for building conversational applications that can leverage external functions and maintain context across interactions. By defining agents and their capabilities, you can create powerful and dynamic chat applications.

Make sure to replace "sk-thekey" with your actual API key when using the OpenAI client in your application.