MCP Bundler

What is MCP Bundler?

MCP Bundler is a library that lets you bundle an MCP server setup easily by composing multiple MCP servers into a single server configuration. It simplifies sharing MCP settings with others.

How to use MCP Bundler?

Import bundler and RequiredEnv from @wrtnlabs/mcp-bundler, then call bundler() with a config object that includes a name, version, and a list of mcpServers. The resulting server can be run via npx example-mcp for stdio mode, npx example-mcp -p <port> for SSE mode, or used directly with InMemoryTransport for InMemory connections.

Key features of MCP Bundler

• Bundles multiple MCP server definitions into one server.
• Supports stdio, SSE, and InMemory transport modes.
• Accepts both inline server objects and external commands.
• Uses RequiredEnv to mark environment variables as mandatory.
• Integrates with existing createServer functions from other MCP packages.

Use cases of MCP Bundler

• Sharing a pre-configured set of MCP servers with teammates.
• Running multiple MCP servers (e.g., Figma, calculator, Notion) through a single entry point.
• Switching between transport modes (stdio, SSE, InMemory) without rewriting server logic.

FAQ from MCP Bundler

What does MCP Bundler do?

It lets you bundle multiple MCP server setup configurations into one cohesive server, making it easier to share and run them together.

How do I run the bundled server in stdio mode?

Execute the bundled server directly, e.g., npx example-mcp in the terminal.

How do I run it in SSE mode?

Use the -p flag with a port number: npx example-mcp -p 4506.

Can I use MCP Bundler with in‑memory transport?

Yes. Use InMemoryTransport.createLinkedPair() to connect a client and server locally without network.

What is RequiredEnv used for?

It marks environment variables that must be provided by the user when the server starts.