M
MCP Directory CURATED DIRECTORY
911 MCP servers indexed
中文
← Back to directory
I

IBM i MCP Server

Official
MCP server for IBM i systems, providing secure SQL-based access.
How to install ↓ Compare GitHub source repository ↗
Category
Database #54 of 58
Stars
★ 85 Popular
Transport
stdio (local process) · Streamable HTTP
Runtime
Node.js 18+ · Docker
Credentials
API key / credential required
License
Apache-2.0
Last commit
Tools
1
42FMRS · D

The IBM i MCP Server provides secure, SQL-based access to IBM i systems via Mapepire, tailored for AI agent interactions with Db2 for i. It is officially maintained but requires the Mapepire prerequisite. With YAML tool definitions and a companion CLI, it is flexible for both AI and terminal users. It is best suited for AI workloads on IBM i needing modern, low-latency database access.

Strongest · Documentation 13/20 Weakest · Reliability 5/20

Reliability
5/20
Security and permissions
6/20
Maintenance
7/20
Documentation
13/20
Setup experience
11/20
Why each score
Reliability 5/20
Only README start commands and architecture are supplied; no package manifest, CI workflow, or tests are visible. The init/tools-list handshake and actual tool behavior cannot be verified, and error handling is unknown. Per the static calibration cap, score is 5.
Security and permissions 6/20
The README shows DB2i_PASS stored via .env and includes DB2i_IGNORE_UNAUTHORIZED=true; YAML tools can execute arbitrary SQL, with no evidence of least privilege, confirmation of dangerous operations, or data-flow disclosure. No malware or real tokens were found, so the red line is not triggered; score is 6.
Maintenance 7/20
The repository is not archived, has an Apache-2.0 license, and shows open issues, but the supplied material lacks commit history, release history, dependency updates, or a security response channel. Maintenance governance evidence is thin; score is 7.
Documentation 13/20
The README is broad: quickstart, environment variables, YAML tool examples, Docker, Python client, deployment, and Mapepire prerequisites, plus external doc links. However, troubleshooting, limits/cost, full parameter schemas, and mainstream client config examples are missing; score is 13.
Setup experience 11/20
Setup steps are clear (npx/Docker, .env, healthz verification), and a Python client example is included. But it depends on an external Mapepire service, lacks mainstream client config JSON (e.g., Claude Desktop), and has no CI/test evidence; score is 11.

Static review · not runListed 2026-08-07

Read the FMRS scoring method →

Fit and risk

What it can accessConnects to a database

Best for

  • Building AI agents or AI-powered applications that interact with IBM i.
  • Requiring remote server support over the MCP protocol (stdio or HTTP).
  • Leveraging Mapepire for fast, lightweight, low-latency database access.
  • Defining custom SQL tools via YAML.

Not for

  • Terminal-first interactions without an MCP client; the `ibmi` CLI is better for that.
  • IBM i systems without Mapepire installed.
  • Scenarios requiring traditional ODBC/JDBC database access.
  • Simple SQL query tools without AI or MCP integration.

Required permissions

  • Requires IBM i credentials (username and password) for database access.
  • Requires Mapepire server running on IBM i (default port 8076).
  • May require network access to the specified port on the IBM i system.
  • Needs access to QSYS2 system views and tables for the example tools.

Risks and side effects

  • Credential management: passing passwords directly can be risky; use environment variables or secure storage.
  • Network attack surface: exposing the MCP server over HTTP could allow unauthorized access unless secured (HTTPS, authentication).
  • Data exposure: SQL tools may access sensitive data; restrict tool scope.
  • Resource consumption: long-running queries may impact IBM i performance.

Setup

Before you start

Runtime：Node.js 18+ · Docker

Other optional settings (1)
MCP_SERVER_CONFIG optional Path to the .env config file when running the Docker container.

Install and configure Mapepire on the IBM i system (default port 8076).
Clone the repository: git clone https://github.com/IBM/ibmi-mcp-server.git and navigate into it.
Configure a .env file with DB2i_HOST, DB2i_USER, DB2i_PASS, DB2i_PORT (default 8076), and DB2i_IGNORE_UNAUTHORIZED.
Start the server with npx -y @ibm/ibmi-mcp-server@latest --transport http --tools ./tools/performance/performance.yaml.
Verify the server health with curl http://localhost:3010/healthz.

Check that it works

Start the server and run curl http://localhost:3010/healthz to confirm it is healthy; the system_status tool should appear in the client's tool list.

Troubleshooting

  1. Verify Mapepire is running on port 8076 and firewall allows inbound connections.
  2. Check that the `.env` file has correct host, user, password, and port settings.
  3. Ensure Node.js 18 or later is installed.
  4. If using Docker, ensure proper volume mounts for the tools directory and `.env` file.
  5. Check server health via `curl http://localhost:3010/healthz`.

Things to try

Once connected, you can ask your AI assistant things like:

  • Show the overall IBM i system status including CPU, memory, and I/O metrics
  • Check current system performance statistics on the IBM i
  • Call the system_status tool to verify system health

Tools 1

system_status read-only
Overall system performance statistics with CPU, memory, and I/O metrics.

Use cases

System monitoring: real-time performance analysis and health checks.
Troubleshooting: diagnose issues using natural language queries.
Reporting: generate system reports and insights.
Automation: execute administrative tasks through conversation.

Supported clients

Claude Desktop
VSCode Copilot
Bob

Listed from the project's documentation, not tested by this site.

Overview

The IBM i MCP Server enables AI agents to interact with IBM i systems through the Model Context Protocol (MCP). It provides secure, SQL-based access to Db2 for i databases via Mapepire, allowing AI applications like Claude, VSCode Copilot, Bob, and custom agents to query system information, monitor performance, and execute database operations. The repo also ships the ibmi CLI, a terminal-first sibling that shares the same YAML-driven SQL tool engine, so the same tool definitions work in both the MCP server and the CLI.

Similar servers

Compare with other MCP servers

Source revision 79ad89a5d2ca Data synced 2026-10-11 Read the FMRS scoring method