ibmi-mcp-server

MCP server for IBM i

📚

• ibmi-mcp-server

  • ✨ Key Features

  • Quick Start

    • 1. Installation

    • 2. Build the Project

    • 3. Create Server .env File

    • 4. Running the Server

    • 5. Run Example Agent

      • Run the example Agent:

      • Run the Example Scripts:

    • 6. Running Tests

  • ⚙️ Configuration

  • IBM i HTTP Authentication (Beta)

    • Authentication Flow

    • Configuration

    • Getting Access Tokens

      • Option 1: Using the Token Script (Recommended)

      • Sequence Overview

    • Client Integration

    • Security Considerations

    • Authentication Endpoints

  • SQL Tool Configuration

    • Sources

    • Tools

    • Toolsets

  • Running the Server (Development)

    • Transport Modes

      • HTTP Transport (Recommended for Development)

      • Stdio Transport (for CLI tools and MCP Inspector)

    • Session Modes (HTTP Only)

    • CLI Options

    • Common Development Scenarios

    • Development Tips

    • Troubleshooting

  • MCP Inspector

  • Docker & Podman Deployment

    • Prerequisites

      • Docker

      • Podman (Alternative to Docker)

      • Build MCP Gateway Image

      • Configure MCP environment

    • Quick Start with Docker

    • Quick Start with Podman

    • Container Architecture

    • 🔧 Service Management

      • Start Services

      • Stop Services

      • View Logs

      • Rebuild Services

    • MCP Gateway UI:

    • Virtual Server Catalog Demo (Comming soon!!)

  • Architecture Overview

  • 🏗️ Project Structure

  • 🧩 Extending the System

    • The "Logic Throws, Handler Catches" Pattern

  • 🌍 Explore More MCP Resources

  • 📜 License

✨ Key Features

Quick Start

1. Installation

Clone the repository and install dependencies:

2. Build the Project

3. Create Server .env File

Fill out the Db2 for i connection details in the .env file:

See more on configuration options in the Configuration section.

4. Running the Server

• Via Stdio (Default):

  npm run start:stdio

• Via Streamable HTTP:

  npm run start:http

  By Default, the server registers SQL tools stored in the prebuiltconfigs directory. This path is set in the .env file (TOOLS_YAML_PATH). You can override the SQL tools path using the CLI:

  • CLI Option: --tools <path>

    npm run start:http -- --tools <path>

  • Transport Options: --transport <type>

    npm run start:http -- --transport http # or stdio

5. Run Example Agent

Make sure that the server is running in http mode:

In another terminal, navigate to the tests/agents directory and follow the setup instructions in the README.

Run the example Agent:

Run the Example Scripts:

6. Running Tests

This template uses Vitest for testing, with a strong emphasis on integration testing to ensure all components work together correctly.

• Run all tests once:

  npm test

• Run tests in watch mode:

  npm run test:watch

• Run tests and generate a coverage report:

  npm run test:coverage

⚙️ Configuration

Configure the server using these environment variables (or a .env file):

To set the server environment variables, create a .env file in the root of this project:

Then edit the .env file with your IBM i connection details.

IBM i HTTP Authentication (Beta)

The server supports IBM i HTTP authentication that allows clients to obtain access tokens for authenticated SQL tool execution. This enables per-user connection pooling and secure access to IBM i resources.

Authentication Flow

1. Client Authentication: Clients authenticate with IBM i credentials via HTTP Basic Auth

2. Token Generation: Server creates a secure Bearer token and establishes a dedicated connection pool

3. Tool Execution: Subsequent tool calls use the Bearer token for authenticated execution

4. Pool Management: Each token maintains its own connection pool for isolation and security

Configuration

To enable IBM i HTTP authentication, we need to set up Encryption keys and configure the server environment. To protect IBM i credentials during transmission, the authentication flow uses RSA and AES encryption. You need to generate an RSA keypair for the server:

Create or update your .env file with the following settings:

Getting Access Tokens

Option 1: Using the Token Script (Recommended)

Use the included get-access-token.js script to obtain authentication tokens:

The script automatically:

• Loads IBM i credentials from .env with CLI fallback

• Fetches the server's public key

• Encrypts credentials client-side

• Requests an access token

• Sets IBMI_MCP_ACCESS_TOKEN environment variable

• Provides copy-paste export commands

Sequence Overview

Client Integration

Once you have a token, use it in your MCP client to authenticate requests:

Security Considerations

Development Environment:

• IBMI_AUTH_ALLOW_HTTP=true allows HTTP for testing

• Use localhost/trusted networks only

• Shorter token lifetimes for testing

Production Environment:

• IBMI_AUTH_ALLOW_HTTP=false enforces HTTPS

• Use proper TLS certificates

• Longer token lifetimes for stability

• Network security and access controls

• Monitor IBMI_AUTH_MAX_CONCURRENT_SESSIONS for resource usage

Authentication Endpoints

When enabled (IBMI_HTTP_AUTH_ENABLED=true), the server provides these endpoints:

SQL Tool Configuration

The Primary way to confgure tools used by this MCP server is through tools.yaml files (see prebuiltconfigs/ for examples). There are 3 main sections to each yaml file: sources, tools, and toolsets. Below is a breakdown of each section

Sources

The sources section of your tools.yaml defines the data sources the MCP server has access to

Tools

The tools section of your tools.yaml defines the actions your agent can take: what kind of tool it is, which source(s) it affects, what parameters it uses, etc.

Toolsets

The toolsets section of your tools.yaml allows you to define groups of tools that you want to be able to load together. This can be useful for defining different sets for different agents or different applications.

More documentation on SQL tools coming soon!

Running the Server (Development)

The server supports multiple transport modes and session configurations for different development scenarios. Use the appropriate startup command based on your needs.

Transport Modes

HTTP Transport (Recommended for Development)

Stdio Transport (for CLI tools and MCP Inspector)

Session Modes (HTTP Only)

The MCP_SESSION_MODE environment variable controls how the HTTP server handles client sessions:

• auto: Automatically detects client capabilities and uses the best session mode

• stateful: Maintains persistent sessions with connection state

• stateless: Each request is independent, no session state maintained

CLI Options

Both transport modes support these command-line options:

Note: CLI arguments override corresponding settings in .env file when provided.

Common Development Scenarios

1. Standard Development Server

2. Custom Tools Path

3. Specific Toolsets Only

Development Tips

• Hot Reloading: Enable YAML_AUTO_RELOAD=true in .env for automatic tool configuration updates

• Verbose Logging: Set MCP_LOG_LEVEL=debug for detailed operation logs

• CORS: Configure MCP_ALLOWED_ORIGINS for web-based clients

• Authentication: Use MCP_AUTH_MODE=ibmi with IBM i HTTP auth for token-based access

Troubleshooting

Port Already in Use

Tools Not Loading

MCP Inspector

The MCP Inspector is a tool for exploring and debugging the MCP server's capabilities. It provides a user-friendly interface for interacting with the server, viewing available tools, and testing queries.

Here are the steps to run the MCP Inspector:

1. Make sure to build the server

   cd ibmi-mcp-server/ npm run build

2. Create an mcp.json file:

   cp template_mcp.json mcp.json

   Fill out the connection details in mcp.json with your IBM i system information. You should use the same credentials as in your .env file:

   { "mcpServers": { "default-server": { "command": "node", "args": ["dist/index.js"], "env": { "TOOLS_YAML_PATH": "prebuiltconfigs", "NODE_OPTIONS": "--no-deprecation", "DB2i_HOST": "<DB2i_HOST>", "DB2i_USER": "<DB2i_USER>", "DB2i_PASS": "<DB2i_PASS>", "DB2i_PORT": "<DB2i_PORT>", "MCP_TRANSPORT_TYPE": "stdio" } } } }

3. Start the MCP Inspector

   npm run mcp-inspector

4. Click on the URL displayed in the terminal to open the MCP Inspector in your web browser.

   Starting MCP inspector... ⚙️ Proxy server listening on 127.0.0.1:6277 🔑 Session token: EXAMPLE_TOKEN Use this token to authenticate requests or set DANGEROUSLY_OMIT_AUTH=true to disable auth 🔗 Open inspector with token pre-filled: http://localhost:6274/?MCP_PROXY_AUTH_TOKEN=EXAMPLE_TOKEN 🔍 MCP Inspector is up and running at http://127.0.0.1:6274 🚀

5. Use the MCP Inspector to explore and test your MCP server's capabilities

   • View available tools and their parameters

   • Test queries against the server

   • Debug issues with tool execution

Docker & Podman Deployment

The project includes a comprehensive docker-compose.yml that sets up the complete MCP gateway with the IBM i MCP Server.

ContextForge MCP Gateway is a feature-rich gateway, proxy and MCP Registry that federates MCP and REST services - unifying discovery, auth, rate-limiting, observability, virtual servers, multi-transport protocols, and an optional Admin UI into one clean endpoint for your AI clients.

Read more about it here.

Prerequisites

Choose one of the following container platforms:

Docker

• Docker Desktop (macOS/Windows): Download here

• Docker Engine (Linux): Installation guide

Podman (Alternative to Docker)

• Podman Desktop (macOS/Windows): Download here

• Podman CLI (Linux): Installation guide

• podman-compose: pip install podman-compose

Build MCP Gateway Image

The docker-compose.yml uses a local build of the MCP Gateway image. To build it, clone the MCP Gateway repository and build the image:

This will create a local image named localhost/mcpgateway/mcpgateway that the docker-compose.yml can use. More details on building the MCP Gateway image can be found in the MCP Gateway Docs.

Configure MCP environment

Create a .env file in the ibmi-mcp-server directory with your IBM i connection details:

make sure to set the follow variables in your .env file:

Note: You need to generate an RSA keypair for the server if you haven't already done so. See the IBM i HTTP Authentication section for instructions.

Once you have your .env file configured, you can start the complete stack using Docker or Podman.

Quick Start with Docker

1. Start the complete stack:

   # Start all services in background docker-compose up -d # Or start specific services docker-compose up -d gateway ibmi-mcp-server postgres redis

2. Verify services are running:

   docker-compose ps

Quick Start with Podman

1. Start the complete stack:

   # Start all services in background podman compose up -d # Or start specific services podman compose up -d gateway ibmi-mcp-server postgres redis

2. Verify services are running:

   podman compose ps

Container Architecture

The docker-compose setup includes these services:

🔧 Service Management

Start Services

Stop Services

View Logs

Rebuild Services

MCP Gateway UI:

After the Containers are up and running, you can access the MCP Context Forge UI at http://localhost:4444

Enter the demo credentials:

• User: admin

• Password: changeme

To Configure the IBM i MCP server is the admin ui, navigate to the "Gateways/MCP Servers" tab. and enter the mcp server endpoint:

• IBM i mcp server endpoint: http://ibmi-mcp-server:3010

Once the MCP server is connect, you can then manage the tools provided by the server:

Virtual Server Catalog Demo (Comming soon!!)

Architecture Overview

This template is built on a set of architectural principles to ensure modularity, testability, and operational clarity.

• Core Server (: The central point where tools and resources are registered. It uses a ManagedMcpServer wrapper to provide enhanced introspection capabilities. It acts the same way as the native McpServer, but with additional features like introspection and enhanced error handling.

• Transports (: The transport layer connects the core server to the outside world. It supports both stdio for direct process communication and a streamable Hono-based http server.

• "Logic Throws, Handler Catches": This is the immutable cornerstone of our error-handling strategy.

  • Core Logic (: This layer is responsible for pure, self-contained business logic. It throws a structured McpError on any failure.

  • Handlers (: This layer interfaces with the server, invokes the core logic, and catches any errors. It is the exclusive location where errors are processed and formatted into a final response.

• Structured, Traceable Operations: Every operation is traced from initiation to completion via a RequestContext that is passed through the entire call stack, ensuring comprehensive and structured logging.

🏗️ Project Structure

• src/mcp-server/: Contains the core MCP server, tools, resources, and transport handlers.

• src/config/: Handles loading and validation of environment variables.

• src/services/: Reusable modules for integrating with external services (DuckDB, OpenRouter).

• src/types-global/: Defines shared TypeScript interfaces and type definitions.

• src/utils/: Core utilities (logging, error handling, security, etc.).

• src/index.ts: The main entry point that initializes and starts the server.

Explore the full structure yourself:

See the current file tree in docs/tree.md or generate it dynamically:

🧩 Extending the System

The template enforces a strict, modular pattern for adding new tools and resources, as mandated by the Architectural Standard. The echoTool (src/mcp-server/tools/echoTool/) serves as the canonical example.

The "Logic Throws, Handler Catches" Pattern

This is the cornerstone of the architecture:

1. logic.ts: This file contains the pure business logic.

   • It defines the Zod schemas for input and output, which serve as the single source of truth for the tool's data contract.

   • The core logic function is pure: it takes validated parameters and a request context, and either returns a result or throws a structured McpError.

   • It never contains try...catch blocks for formatting a final response.

2. registration.ts: This file is the "handler" that connects the logic to the MCP server.

   • It imports the schemas and logic function from logic.ts.

   • It calls server.registerTool(), providing the tool's metadata and the runtime handler.

   • The runtime handler always wraps the call to the logic function in a try...catch block. This is the only place where errors are caught, processed by the ErrorHandler, and formatted into a standardized error response.

This pattern ensures that core logic remains decoupled, pure, and easily testable, while the registration layer handles all transport-level concerns, side effects, and response formatting.

🌍 Explore More MCP Resources

Looking for more examples, guides, and pre-built MCP servers? Check out the companion repository:

➡️ cyanheads/model-context-protocol-resources

📜 License

This project is licensed under the Apache License 2.0. See the LICENSE file for details.