https://github.com/vgulerianb/agents-js
A versatile JavaScript library for designing and managing intelligent agents capable of interacting with AI-driven chat models, leveraging OpenAI's GPT for dynamic conversations and task automation.",
https://github.com/vgulerianb/agents-js
agents ai multiagent-systems openai swarm
Last synced: 5 months ago
JSON representation
A versatile JavaScript library for designing and managing intelligent agents capable of interacting with AI-driven chat models, leveraging OpenAI's GPT for dynamic conversations and task automation.",
- Host: GitHub
- URL: https://github.com/vgulerianb/agents-js
- Owner: vgulerianb
- Created: 2024-11-09T14:27:19.000Z (6 months ago)
- Default Branch: main
- Last Pushed: 2024-11-15T16:24:40.000Z (5 months ago)
- Last Synced: 2024-11-18T09:46:05.551Z (5 months ago)
- Topics: agents, ai, multiagent-systems, openai, swarm
- Language: JavaScript
- Homepage:
- Size: 12.7 KB
- Stars: 1
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
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:
```bash
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
```javascript
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.
```javascript
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);
})();
```
> If you want to use TuneStudio you can pass base url to the client like this:
>
> ```javascript
> const client = new OpenAI({
> apiKey: "sk-tune-key",
> baseURL: "https://proxy.tune.app",
> });
> ```
````
### Example 3: Handling Tool Calls
You can define functions that the agent can call during its operation.
```javascript
const { AgentController, Agent, Result } = require("agents-js");
// Define a simple function that the agent can use
function add({ a, b }) {
/**
* @param {number} a - First number
* @param {number} b - Second number
* @description Adds two numbers
*/
return new Result({ value: `${a + b}` });
}
// 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 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.
```javascript
const { AgentController, Agent, Result } = require("agents-js");
// Define a function using context variables
function greet({ name }) {
/**
* @param {string} name - Name of the person to greet
* @description Greets a person by name
*/
return new Result({ value: `Hello, ${name}!` });
}
// 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 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.