Auto Park AI Agent

Collection of Simple Agents using LangChain with Anthropic Claude and Google Gemini

Overview

This project demonstrates basic agent implementations using the LangChain framework with two major LLM providers:

• Anthropic Claude (with tool calling capabilities)
• Google Gemini

Features

✅ Anthropic Claude 3.5 Sonnet agent with function calling ✅ Custom weather tool implementation ✅ Google Gemini 2.5 Flash integration ✅ TypeScript support ✅ Built with official LangChain libraries

Installation

# Install dependencies
pnpm install

Configuration

1. Copy the example environment file:

cp .env.example .env

2. Add your API keys in the .env file:

ANTHROPIC_API_KEY=your_anthropic_key_here
GOOGLE_API_KEY=your_google_api_key_here

Usage

Run Anthropic Claude Agent

# Production mode
pnpm start

# Development mode (with hot reload)
pnpm dev

Run Google Gemini Agent

pnpm start:gemini

Available Scripts

Command	Description
pnpm build	Compile TypeScript to JavaScript
pnpm start	Build and run Anthropic agent
pnpm start:gemini	Build and run Gemini agent
pnpm dev	Run Anthropic agent in development mode
pnpm type-check	Validate TypeScript types

Project Structure

langchain_agents/
├── app/
│   ├── main.ts          # Anthropic Claude agent implementation
│   └── gemini/
│       └── index.ts     # Google Gemini implementation
├── dist/                # Compiled output (auto-generated)
├── .env.example         # Environment variables template
├── package.json
├── tsconfig.json
└── README.md

Dependencies

• @langchain/anthropic - Anthropic integration
• @langchain/google - Google Gemini integration
• langchain - Core LangChain framework
• zod - Schema validation for tools
• typescript - TypeScript compiler
• dotenv - Environment variable loading