Kintone Documentation MCP Server

The Introduction of Kintone Documentation MCP Server

This article introduces the features and usage of Kintone's official Documentation MCP server.

What is the Kintone Documentation MCP Server

The Kintone Documentation MCP Server is a tool that lets AI retrieve the latest official Kintone documentation and API information.

In contrast to the Kintone MCP Server , this server does not perform any operations against an actual Kintone environment.
It's well suited to use cases such as:

Wanting to always reference the latest official documentation, regardless of the training cutoff of the AI Platform or Model the user is using.
Wanting responses that reflect an understanding of best practices and coding guidelines.
Wanting to develop with AI while keeping token consumption lower than when using a generative AI platform's built-in search tools.

Quickstart

The Kintone Documentation MCP Server is offered in two forms: a remote server version and a packaged version.
Both provide the same behavior and functionality. However, the package version needs to be updated manually.

Remote server version

Users can use the Kintone Documentation MCP Server by pointing their generative AI tool at https://mcp.kintone.dev/mcp.

For Claude

In Claude's Web or Desktop app, select "Customize" in the left panel. Choose "Connectors", click the "+" button, and select "Add custom connector".

Enter https://mcp.kintone.dev/mcp in the connector's URL field.

Click "Add" and confirm it has been added.

For Claude Code

Run the following command:

1

  claude mcp add --transport http kintone-docs https://mcp.kintone.dev/mcp

Optionally, use the --scope option to specify the configuration's scope.
As in the command above, if omitted, it will be added with local scope (valid only within the current project).
For more details on configuration, see the official Claude Code documentation .

For ChatGPT

From ChatGPT's left menu, click "Plugins" and then the + (Add Connector) button.

Enter a "Name" and "Description", and fill in https://mcp.kintone.dev/mcp for the "Connect" settings.
Select "No authentication" for the "Authentication" settings.

Click "Connect" and confirm it has been added.

For Cursor

Add the following to ~/.cursor/mcp.json (global settings) or .cursor/mcp.json (project settings).

1
2
3
4
5
6
7


  {
    "mcpServers": {
      "kintone-docs": {
        "url": "https://mcp.kintone.dev/mcp"
      }
    }
  }

Warning

The AI tools introduced here are only examples.
Refer to each product's official documentation for setup instructions.
Depending on the generative AI tool's security settings, users may not be able to add a remote MCP server.
Administrator approval or configuration changes may be required in some cases.

Package version

For Claude users, the MCP Server can also be installed from the following MCPB file.
Refer to the link below for installation steps.
Kintone MCP Server installation instructions

Download the package version ()

Although the installation format is MCPB, processing itself is performed on the remote server.
The packaged version does not have an auto-update feature. Users will need to self-check whether they are using the latest version.

Request for feedback

The Kintone Documentation MCP Server is a tool aimed at improving the Kintone development experience with generative AI.
We'd love to hear feedback from everyone who uses AI when developing on Kintone on a daily basis. (Estimated time: about 3 minutes)

Send feedback

Release objectives and goals

As development using generative AI becomes more common, there is growing demand for providing Kintone documentation in a format that's easy for AI to process.
The Kintone Documentation MCP Server is provided to help solve the challenges developers face when using generative AI.

Solved pain points

Generative AI's training data only includes information up to its training cutoff. It can sometimes fail to accurately reflect the specifications of recently added APIs.
As a result, it may generate code that uses APIs that don't exist, or that are deprecated.
It can also produce implementations that violate Kintone-specific development conventions and best practices, or code that simply doesn't work.
While one workaround is to paste large amounts of documentation into the context every time, this raises concerns about token consumption.

Usage cautions

The Kintone Documentation MCP Server is a tool used by generative AI models.
Please keep the following points in mind when using it.

Responses and how the MCP tools are used will vary depending on the model the user is using.
Even when referencing the latest documentation, there's no guarantee of how the generative AI will make use of it. Users may need to guide it appropriately through their prompts.
Because the tools may be called repeatedly, token usage may increase.
- To prevent excessive calls, the server includes instructions for the AI on how to use the tools.
- If users are concerned about token usage, they can disable the Kintone Documentation MCP Server in their AI Platform temporarily.
Depending on the model and settings, the tools may not be used at all.
In that case, explicitly stating in the prompt such phrases as "please look this up using the Kintone Documentation MCP Server tool" will often resolve the issue.

Roadmap

The following tools are currently available.

search_docs
Searches the JavaScript API and REST API Documentation.
search_resources
Searches articles related to integrations, SDKs, best practices, authentication, etc.
get_page
Retrieves a specific documentation page by its path.

As of July 2026, the following feature additions are under consideration.

check_updates
This tool works by giving a customization or plugin code. The tool will check the Kintone API's update history, report whether any relevant updates apply, and suggest fixes.
plan_app
This would propose a design for a Kintone app.

We hope that combining these features with the Kintone MCP Server will enable more advanced AI agents for Kintone development.

Kintone MCP Server

AI