@uaito/ai

This package provides a powerful Agent class that can be used to orchestrate tasks with various Large Language Models. It also includes a command-line interface (CLI) for interacting with the agent directly from your terminal.

Installation

npm install @uaito/ai

or

yarn add @uaito/ai

Library Usage

The main export of this package is the Agent class. You can use it to wrap any LLM provider that conforms to the @uaito/sdk's BaseLLM interface.

import { Agent } from '@uaito/ai';
import { OpenAI } from '@uaito/openai';
import { MessageArray } from '@uaito/sdk';

async function main() {
  const llm = new OpenAI({
    options: {
        // ... your OpenAI options
    }
  });
  
  const agent = new Agent(llm);
  
  const history = new MessageArray([]); // or provide existing message history
  await agent.addInputs(history);
  
  const { response } = await agent.performTask("Hello, who are you?");
  
  for await (const chunk of response) {
    if (chunk.type === 'message') {
        for (const content of chunk.content) {
            if (content.type === 'text') {
                process.stdout.write(content.text);
            }
        }
    }
  }
}

main();

CLI Usage

This package also provides the uaito-cli command, which allows you to run the agent from your command line.

Command

uaito-cli run <message> [options]

Runs the application with a given configuration.

Arguments

• message: (Required) The message to send to the agent.

Options

• --provider, -p: (Required) The LLM provider to use.
  • Choices: OpenAI, Anthropic, Grok, Google, Local, API
• --model, -m: (Required) The model to use. The available models depend on the selected provider. See provider-specific details below.
• --apiKey: The API key for the provider. If not provided, the CLI will look for an environment variable.

Provider-specific Details and Examples

OpenAI

• Provider value: OpenAI
• Models: gpt-5-nano, gpt-5-mini.
• API Key: Use the --apiKey option or set the OPENAI_API_KEY environment variable.

Example:

npx uaito-cli run "Translate 'Hello, world!' to French." -p OpenAI -m gpt-5-nano

Anthropic

• Provider value: Anthropic
• Models: Any model string supported by the Anthropic API. E.g., claude-4-sonnet (which corresponds to "claude-sonnet-4-2025-05-14").
• API Key: Use the --apiKey option or set the ANTHROPIC_API_KEY environment variable.

Example:

npx uaito-cli run "What is the capital of France?" -p Anthropic -m claude-4-sonnet

Grok

• Provider value: Grok
• Models: grok-4.
• API Key: Use the --apiKey option or set the GROK_API_KEY environment variable.

Example:

npx uaito-cli run "Explain quantum computing in simple terms." -p Grok -m grok-4

Google

• Provider value: Google
• Models: gemini-2.5-pro.
• API Key: Use the --apiKey option or set the GOOGLE_API_KEY environment variable.

Example:

npx uaito-cli run "What is the capital of Spain?" -p Google -m gemini-2.5-pro

Local

• Provider value: Local
• Models: Only onnx-community/Lucy-ONNX is currently supported.
• API Key: Not required.

Example:

npx uaito-cli run "Write a short poem about coding." -p Local -m onnx-community/Lucy-ONNX

API

• Provider value: API
• Description: This provider uses the UAITO API endpoint. It currently uses the Anthropic provider under the hood.
• Models: Any model string supported by the backend.
• API Key: Use the --apiKey option or set the GROK_API_KEY environment variable.

Example:

npx uaito-cli run "Summarize the provided text." -p API -m some-anthropic-model