Maintenance Connection MCP server

Available tools & resources for
Maintenance Connection

search_banks

Find banks by name or identifier. Returns the financial institutions on file.

update_asset

Update an existing asset. Change its location, status, attributes, or identifiers.

create_asset

search_assets

Find assets by name, type, location, or status. Returns identifiers, location, and maintenance history references.

search_addresses

Find addresses by entity or type. Returns billing, shipping, and site addresses linked to records.

update_account

Update an existing account. Change its name, type, or status.

create_account

Create a new general ledger account. Set its number, name, and type.

search_accounts

Find general ledger accounts by name, number, or type. Returns the chart of accounts and balances.

get_raw_data_by_id

Fetch a single native record by its ID in the source system, in the system's own format. Pairs with search_raw_data for custom and platform-specific objects.

search_raw_data

Query a system's native entities and fields directly, beyond the unified data model. Reach custom objects and platform-specific data by their original names.

get_connection_details

Return the details of a connection, including which system it points to and the objects and fields available on it. Use it to discover what an agent can do before making other calls.

...

ONE MCP. EVERY INDUSTRIAL PRODUCT.

Maintenance Connection integration your roadmap keeps pushing. Already built.

Connect once over MCP. Your agent reads data and triggers actions in Maintenance Connection, plus 2,000 other systems, without you building or maintaining a single integration.

Your agent runs Maintenance Connection

One MCP endpoint and your agent can pull data and act on Maintenance Connection. The same connection reaches 2,000+ other products behind it.

Build the agent once

Your agent gets Maintenance Connection's real objects and actions, not a stripped-down common denominator. Work orders stay work orders, assets stay assets. Add SAP or Maximo through the same server when you need them.

AI Agent only does what it's allowed to

Each connection runs on its own scoped credentials, so the agent inherits your access model. Every call is logged, so you can see exactly what it did.

Live in three steps

Makini Link - Connect

670e44721718a798d6dc119c_2koYvrO4V6RzNU3StEWkLQQqyUoyugIW_McFq-GigTo.png

Connect to

Maintenance Connection

credentials

User name

••••••••

Connect

Maintenance Connection

→

mcp.makini.io - Endpoint

MCP ENDPOINT

https://mcp.makini.io/UNIQUE_ID

This unique URL is your key - no separate API token to manage.

SCOPED TOKEN

Maintenance Connection

· scoped to your login

● Active

RECENT CALLS

search_work_orders

200

create_work_order

201

search_inventory_transfers

200

Agent config - MCP server

Open claude_desktop_config.json and add the snippet under "mcpServers" (create the file if it doesn't exist):

macOS ~/Library/Application Support/Claude/claude_desktop_config.json

Windows %APPDATA%\Claude\claude_desktop_config.json

Linux ~/.config/Claude/claude_desktop_config.json

{
  "mcpServers": {
    "makini": {
      "command": "npx",
      "args": ["-y", "mcp-remote@latest",
        "https://mcp.makini.io/UNIQUE_ID"]
    }
  }
}

Save the file and restart Claude Desktop.
Prefer the CLI? Claude Code:
claude mcp add --transport http makini https://mcp.makini.io/UNIQUE_ID
Add --scope project to save the config to .mcp.json in the current directory only.

Add the snippet under "mcpServers" (create the file if it doesn't exist), then restart Cursor:

macOS / Linux ~/.cursor/mcp.json
Windows %APPDATA%\Cursor\mcp.json
Project-only .cursor/mcp.json in your project root

{
  "mcpServers": {
    "makini": {
      "command": "npx",
      "args": ["-y", "mcp-remote@latest",
        "https://mcp.makini.io/UNIQUE_ID"]
    }
  }
}

VS Code uses "servers" as the root key - not "mcpServers". It applies the change automatically:

Global Command Palette → MCP: Open User Configuration
Project-only create .vscode/mcp.json in your workspace root

{
  "servers": {
    "makini": {
      "type": "http",
      "url": "https://mcp.makini.io/UNIQUE_ID"
    }
  }
}

Add the snippet under "mcpServers" (create the file if it doesn't exist), then reload the Windsurf window:

macOS / Linux ~/.codeium/windsurf/mcp_config.json
Windows %APPDATA%\Codeium\windsurf\mcp_config.json

{
  "mcpServers": {
    "makini": {
      "command": "npx",
      "args": ["-y", "mcp-remote@latest",
        "https://mcp.makini.io/UNIQUE_ID"]
    }
  }
}

Add the snippet under "mcpServers" (create the file if it doesn't exist). Cline picks up the change automatically:

macOS / Linux ~/.vscode/globalStorage/saoudrizwan.claude-dev/settings/cline_mcp_settings.json
Windows %APPDATA%\Code\User\globalStorage\saoudrizwan.claude-dev\settings\cline_mcp_settings.json

{
  "mcpServers": {
    "makini": {
      "url": "https://mcp.makini.io/UNIQUE_ID",
      "disabled": false,
      "autoApprove": []
    }
  }
}

Add the makini entry under "context_servers" in settings.json. Zed picks up the change automatically:

macOS ~/Library/Application Support/Zed/settings.json
Windows %APPDATA%\Zed\settings.json
Linux ~/.config/zed/settings.json

{
  "context_servers": {
    "makini": {
      "url": "https://mcp.makini.io/UNIQUE_ID"
    }
  }
}

WHY MAKINI MCP

Everything your agent needs to act on industrial systems

The hard part of an agent isn't the model. It's giving it safe, reliable access to the systems that actually run the business. Makini handles that part. Connect once over MCP and your agent reads and acts on Maintenance Connection plus 2,000 other systems, with auth, scoping, and logging built in.

Native tools, not a lowest common denominator

Your agent calls the same four tools, search, get, create, update, across all 2,000 systems. It runs real endpoints, objects, and filters, exposed in real time. Custom fields and custom objects included. No abstract data model to learn, and nothing lost in translation.

One toolset

search get create update

Maintenance Connection IBM Maximo Oracle WMS Infor WMS Blue Yonder +2,000

Reach systems that never had an API

Plenty of industrial software never shipped a REST API or an MCP server, including 20-year-old on-prem systems (SAP R/3, MP2, ABS NS, CorrigoPro, etc.). Your agent reaches them anyway. Almost no other MCP provider can.

On-prem · 2003

No API

AI agent

over MCP

Connect by logging in

Pick a product and sign in with the credentials you already use, whether it's you, your code, or an agent over MCP. The same login flow works everywhere.

User name

••••••••••

✓Connected to Maintenance Connection

Scoped, logged, accounted for

Each connection runs on its own token and can only do what those credentials allow in the source system. Every call the agent makes is logged and searchable, so you can see exactly what it did.

mcp.makini.io/UNIQUE_ID scoped

09:42:01search_work_orders200

09:42:04create_work_order201

09:42:09update_stock_level200

Works with any MCP client

Point Claude, ChatGPT, Cursor, or n8n at your endpoint with one config block. No custom adapters, no lock-in, switch clients anytime.

Claude ChatGPT Cursor VS Code Windsurf Cline n8n

mcp.makini.io/v1 1 endpoint

Connections that don't silently break

Makini monitors every tool call and alerts you the moment one fails, with the error detail to fix it. Your agent is never quietly calling a tool that went dead.

Sync failed on Maintenance Connection

503 upstream timeout - alert sent

IBM Maximohealthy

Oracle WMShealthy

Infor WMShealthy

Security and governance

Control what the agent touches. See everything it does.

Writing to a live CMMS or ERP is where teams get nervous, and they are right to be. Makini hands you the controls. Every connection is scoped to what its credentials can do, every call the agent makes is logged. Highest security rating from the most demanding in the business. SOC 2 Type II, UK Cyber Essentials, and GDPR compliant.

Scoped, revocable access

Each connection runs on its own token, scoped to what its credentials can do in the source system. Revoke it and access is cut off immediately.

A log of every action

Every call is logged and searchable, from API requests to workflow executions. Stream the audit trail into Sentry, a webhook, or syslog, so it sits where your team already looks

Independently verified

SOC 2 Type II, UK Cyber Essentials, and GDPR compliant with independent tests twice a year.

Run it where compliance demandsn

In-region hosting for the EU, UK, Canada, and Australia, a dedicated single-tenant instance, or self-hosted entirely in your own environment. GovCloud and ITAR are supported and running in production today.