examples

LangChainJS-MCP-Adapters Examples

This directory contains examples demonstrating how to use the @langchain/mcp-adapters library with various MCP servers

Running the Examples

# Build all examples
yarn build:examples

# Run specific example
cd examples && npx -y tsx firecrawl_custom_config_example.ts

Example Descriptions

Filesystem LangGraph Example (filesystem_langgraph_example.ts)

Demonstrates using the Filesystem MCP server with LangGraph to create a structured workflow for complex file operations. The example creates a graph-based agent that can perform various file operations like creating multiple files, reading files, creating directory structures, and organizing files.

Firecrawl - Custom Configuration (firecrawl_custom_config_example.ts)

Shows how to initialize the Firecrawl MCP server with a custom configuration. The example sets up a connection to Firecrawl using SSE transport, loads tools from the server, and creates a React agent to perform web scraping tasks and find news about artificial intelligence.

Firecrawl - Multiple Servers (firecrawl_multiple_servers_example.ts)

Demonstrates how to use multiple MCP servers simultaneously by configuring both Firecrawl for web scraping and a Math server for calculations. The example creates a React agent that can use tools from both servers to answer queries involving both math calculations and web content retrieval.

LangGraph - Complex Config (langgraph_complex_config_example.ts)

Illustrates using different configuration files to set up connections to MCP servers, with a focus on the Math server. This example shows how to parse JSON configuration files, connect to a Math server directly, and create a LangGraph workflow that can perform mathematical operations using MCP tools.

LangGraph - Simple Config (langgraph_example.ts)

Shows a straightforward integration of LangGraph with MCP tools, creating a flexible agent workflow. The example demonstrates how to set up a graph-based structure with separate nodes for LLM reasoning and tool execution, with conditional routing between nodes based on whether tool calls are needed.

Launching a Containerized MCP Server (mcp_over_docker_example.ts)

Shows how to run an MCP server inside a Docker container. This example configures a connection to a containerized Filesystem MCP server with appropriate volume mounting, demonstrating how to use Docker to isolate and run MCP servers while still allowing file operations.

Requirements

Ensure you have the correct environment variables set in your .env file:

OPENAI_API_KEY=your_openai_api_key
FIRECRAWL_API_KEY=your_firecrawl_api_key
OPENAI_MODEL_NAME=gpt-4o  # or your preferred model