Jaou Ensatsu Kokuryu

Provides read-only access to FileMaker data and metadata via MCP for cross-layout analysis and global searches.

typescript

GitHub Stars

typescript

Language

6 months ago

First Indexed

3 months ago

Catalog Refreshed

Documentation & install

Readme and setup notes from the catalogue, plus a client-ready config you can copy for your MCP host.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs

{
  "mcpServers": {
    "ta-toshio-filemaker-mcp": {
      "command": "claude",
      "args": [
        "mcp",
        "add",
        "filemaker",
        "--transport",
        "stdio",
        "--scope",
        "project",
        "--env",
        "FM_SERVER=https://your-server.com",
        "--env",
        "FM_DATABASE=your-database",
        "--env",
        "FM_USERNAME=your-username",
        "--env",
        "FM_PASSWORD=your-password",
        "--",
        "node",
        "/path/to/jaou-ensatsu-kokuryu-filemaker-mcp/dist/index.js"
      ],
      "env": {
        "FM_SERVER": "https://your-server.com",
        "FM_DATABASE": "your-database",
        "FM_PASSWORD": "your-password",
        "FM_USERNAME": "your-username"
      }
    }
  }
}

You use this MCP server to analyze FileMaker data via the Data API, extract metadata, infer relationships, and perform cross-layout searches in a read-only fashion. It helps you quickly inspect layouts, fields, scripts, and sample data across your FileMaker databases while keeping security and data integrity in mind.

How to use

You connect your MCP client (such as Claude Code, Cursor, or Codex) to the FileMaker MCP server to run analysis and discovery tasks. Start by authenticating if required, then issue high-level tasks like retrieving layout metadata, exporting database metadata, listing fields, or performing a global search across layouts. All actions are read-only, so you can safely explore without modifying data.

How to install

Prerequisites: you need Node.js installed on your machine and a working MCP client that supports HTTP or STDIO transport. Install the repository dependencies and build the MCP server artifacts.

pnpm install
pnpm run build

Additional configuration and usage notes

To use the FileMaker MCP server with your AI editors, set up an MCP connection for each editor. You will provide the FileMaker server URL, database name, and credentials so the editor can request metadata and read data through the MCP server.

Environment variables you commonly configure include the server URL, database, username, and password. Make sure to keep passwords out of logs and avoid writing sensitive data to disk.

You can initiate the MCP connection for each editor via the included command-line flow or by editing the editor’s local configuration. The recommended method is to run a dedicated MCP add command for your editor, then start the MCP server runtime so your editor can load the tool.

Security considerations

The MCP server is designed to be read-only to protect data integrity. Sensitive credentials should not be logged, and file system writes should be disabled to mitigate risk. Use secure transport (HTTPS) for the FileMaker server URL and rely on strong authentication practices for the editor connections.

Examples of common tasks

Fetch the layout metadata for a database to understand fields and script usage. Export the full database metadata to inspect all layouts, fields, and scripts. Perform a global search to locate records across multiple layouts based on specific field values.

Available tools

fm_login

fm_logout

End the current FileMaker session.

fm_validate_session

Check whether the current session is still valid.

fm_get_layouts

Retrieve a list of layouts available in the database.

fm_get_layout_metadata

Get field definitions and portal information for a specific layout.

fm_get_scripts

List scripts defined in the FileMaker solution.

fm_list_value_lists

List value lists used in the database.

fm_get_records

Retrieve records from a layout with optional paging.

fm_get_record_by_id

Retrieve a single record by its internal ID.

fm_find_records

Search for records using a query across fields.

fm_get_record_count

Return the number of records in a layout that match a query.

fm_export_database_metadata

Export a complete snapshot of layouts, fields, scripts, and value lists.

fm_infer_relationships

Infer potential relationships from field name patterns and metadata.

fm_analyze_portal_data

Analyze portal structures and related records.

fm_global_search_data

Search across multiple layouts for matching values.

fm_global_search_fields

Search across field metadata for name patterns.

Built by

VeilStrat

AI signals for GTM teams