Kintone

You can run a Python-based MCP Server that connects to kintone to read and manipulate data. It supports secure authentication, full CRUD operations, pagination, advanced queries, file handling, comments, status workflows, asynchronous processing, robust error handling, and internationalization, making it practical for AI assistants to work with kintone data.

How to use

Use an MCP client to connect to the kintone MCP Server and perform actions such as retrieving records, creating new ones, updating fields, uploading files, adding comments, and updating statuses. You can work with single operations or batch operations to manage many records efficiently. Start by choosing a connection method that fits your environment—remote HTTP access or local stdio execution via uvx.

How to install

Prerequisites: Python 3.12 or newer, an MCP-capable client, and access credentials to your kintone environment.

Option A: Run the MCP server locally using uvx (recommended for quick start) and a Git-based source.

{
  "mcpServers": {
    "kintone": {
      "command": "uvx",
      "args": [
        "--from",
        "git+https://github.com/r3-yamauchi/kintone-mcp-server-python3.git",
        "kintone-mcp-server-python3"
      ],
      "env": {
        "KINTONE_DOMAIN": "your-subdomain.cybozu.com",
        "KINTONE_USERNAME": "your-username",
        "KINTONE_PASSWORD": "your-password"
      }
    }
  }
}

Option B: Run the MCP server locally from source code in a Python environment

Clone the project, set up a virtual environment, install development dependencies, and start the server.

# Clone the repository
git clone https://github.com/r3-yamauchi/kintone-mcp-server-python3.git
cd kintone-mcp-server-python3

# Create and activate a virtual environment
python -m venv venv
source venv/bin/activate  # macOS/Linux
# venv\Scripts\activate  # Windows

# Install development dependencies
pip install -e ".[dev]"

# Set environment variables in advance or use a .env file
# Example run (start server in module form)
python -m kintone_mcp_server_python3

Additional usage notes

Two practical, ready-to-use configurations are shown for running the server via uvx. The first uses user credentials to access kintone, and the second demonstrates a variant suitable for integration with editors or IDEs. Ensure you replace domain and credentials with your actual values before starting.

Troubleshooting and tips

Common issues include connection errors, authentication failures, and permission problems. Verify that the domain is correct, your network allows access to kintone, and that your API token or user credentials have the necessary permissions for the target apps. If you encounter authentication errors, confirm that the credentials you provided are valid and that API tokens have the required scopes.

For debugging, you can enable verbose logs to identify where the process is failing. Ensure you restart the MCP client after changing configuration values.

Examples of supported actions (high level)

The MCP server exposes tools for record management, comments and status handling, file management, and app information access. Use the MCP client to call these actions in your workflow against your kintone environment.

Available tools

get_record

Retrieve a single record by app ID and record ID to obtain detailed field values.

get_records

Fetch a list of records with optional filtering, including pagination.

get_all_records

Retrieve all records from an app, automatically handling pagination.

add_record

Create a new record in an app with specified field values.

add_records

Add multiple records in a single operation (batch, up to 100).

update_record

Update a single record with new field values.

update_records

Update multiple records in a batch (up to 100).

get_comments

Retrieve comments on a specific record.

add_comment

Add a comment to a specific record, with optional mentions.

update_status

Update the status of a specific record in a workflow.

update_statuses

Batch update statuses for multiple records.

upload_file

Upload a file to kintone and obtain a file key.

download_file

Download a file from kintone using a file key.

get_app

Retrieve details about a specific kintone app.

get_apps

Search and retrieve information about apps in the space.

get_form_fields

Fetch the form field definitions for an app, with optional language localization.