README.md

Remote API Agent Toolkit for TypeScript

This package contains the TypeScript implementation of the Remote AI Agent Toolkit. It provides tools for integrating with the Remote API in agentic workflows.

Installation

npm install @remoteoss/ai-agent-toolkit

Requirements

• Node.js 18+
• A Remote API Key

Usage

The library needs to be configured with your Remote API key.

LangChain Toolkit

To use the toolkit with LangChain, you can instantiate RemoteApiAgentToolkit and pass the tools to your agent.

import { RemoteApiAgentToolkit } from '@remoteoss/ai-agent-toolkit/langchain';
import { ChatOpenAI } from '@langchain/openai';
import { AgentExecutor, createStructuredChatAgent } from 'langchain/agents';
import { hub } from 'langchain/hub';

// Initialize the toolkit
const toolkit = new RemoteApiAgentToolkit({
  apiKey: process.env.REMOTE_API_KEY!,
});

const tools = toolkit.getTools();

// Initialize the LLM and agent
// This example uses OpenAI, but it's compatible with other models
const llm = new ChatOpenAI({ modelName: 'gpt-4o', temperature: 0 });
const prompt = await hub.pull<any>('hwchase17/structured-chat-agent');

const agent = await createStructuredChatAgent({
  llm,
  tools,
  prompt,
});

const agentExecutor = new AgentExecutor({
  agent,
  tools,
});

const result = await agentExecutor.invoke({
  input: 'Can you list the current employments?',
});

console.log(result);

Tool Configuration

You can configure the tools available to the agent by passing the allowedTools array in the toolkit's constructor.

const toolkit = new RemoteApiAgentToolkit({
  apiKey: process.env.REMOTE_API_KEY!,
  allowedTools: ['list_employments', 'create_time_off'],
});

Model Context Protocol (MCP) Server

If you have cloned the repository, you can run the MCP server directly.

First, install the dependencies:

cd mcp
npm install

Then, run the server:

npm run start -- --api-key=YOUR_REMOTE_API_KEY

You can also set the REMOTE_API_KEY and ALLOWED_TOOLS environment variables instead of passing them as arguments.