gcs-mcp-server by Utopiality7 - MCP Server

MyEnhancedGCSMCPServer

A fastmcp server that exposes a comprehensive set of tools for managing Google Cloud Storage (GCS). This server allows you to interact with GCS buckets and objects through a simple HTTP API.

Features

This server provides a wide range of tools to manage GCS resources:

Bucket Management: Create, delete, list, and get metadata for GCS buckets.
Object Management: Upload, download, delete, list, rename, and copy objects (blobs).
Metadata: Retrieve detailed metadata for both buckets and objects.
Security: Generate temporary, secure signed URLs for object access.

Prerequisites

Python 3.8+
uv installed (pip install uv).
A Google Cloud Platform (GCP) project with the Cloud Storage API enabled.
The Google Cloud SDK installed and configured on your local machine.

Setup and Installation

Clone the Repository

git clone <your-repository-url>
cd gcs-mcp-server

Create Virtual Environment and Install Dependencies uv can handle both steps in one go. It will create a virtual environment in the .venv directory and install all packages from requirements.txt.
```
uv venv
uv pip install -r requirements.txt
```

Authentication

To allow the server to access your Google Cloud resources, you need to provide application default credentials. The easiest way to do this for local development is with the gcloud CLI:

gcloud auth application-default login

This command will open a browser window for you to log in and grant the necessary permissions.

Running the Server

The best way to run the server during development is with the fastmcp run command. It starts your server and launches the MCP Inspector, a web UI for interactively testing your tools.


fastmcp run main.py --transport="http" --port="8080"
fastmcp dev main.py --ui-port="9080" --server-port="5080"

You should see output confirming that the server is running and listing the available tools:

================================================================================
MyEnhancedGCSMCPServer is running on http://127.0.0.1:8080/mcp
The following tools are available:
- greet(name: str) -> str
- list_gcs_buckets() -> list[str]
- ... (and all other tools)
================================================================================

All tool endpoints are available under the /mcp/ path. You can interact with them using any HTTP client, such as curl. You can interact with the server's tools using the mcp-inspector command-line utility, which comes bundled with fastmcp. It provides a more convenient way to call tools than using curl.

The uv run command ensures that mcp-inspector is executed within your project's virtual environment.

API Reference: Available Tools

Below is a complete list of the available tools and their parameters.

Tool	Description	Parameters
`greet`	Returns a friendly greeting.	`name: str`
`list_gcs_buckets`	Lists all GCS buckets.	(None)
`create_bucket`	Creates a new GCS bucket.	`bucket_name: str`, `location: str = "US"`
`delete_bucket`	Deletes a GCS bucket.	`bucket_name: str`
`list_objects`	Lists all objects in a bucket.	`bucket_name: str`
`upload_blob`	Uploads a file to a bucket.	`bucket_name: str`, `source_file_name: str`, `destination_blob_name: str`
`download_blob`	Downloads a blob from a bucket.	`bucket_name: str`, `blob_name: str`, `destination_file_name: str`
`delete_blob`	Deletes a blob from a bucket.	`bucket_name: str`, `blob_name: str`
`get_bucket_metadata`	Retrieves metadata for a bucket.	`bucket_name: str`
`get_blob_metadata`	Retrieves metadata for an object.	`bucket_name: str`, `blob_name: str`
`generate_signed_url`	Generates a temporary signed URL.	`bucket_name: str`, `blob_name: str`, `expiration_minutes: int = 15`
`rename_blob`	Renames/moves an object.	`bucket_name: str`, `blob_name: str`, `new_name: str`
`copy_blob`	Copies an object to another bucket.	`source_bucket_name: str`, `blob_name: str`, `destination_bucket_name: str`, `destination_blob_name: str`
`set_bucket_cors`	Sets the CORS configuration for a bucket.	`bucket_name: str`, `cors_rules: list[dict]`

Example:

from google.api_core import exceptions

@mcp.tool
def delete_bucket(bucket_name: str) -> str:
    """Deletes a GCS bucket"""
    try:
        bucket = storage_client.bucket(bucket_name)
        bucket.delete(force=True)
        return f"Bucket '{bucket_name}' deleted successfully."
    except exceptions.NotFound:
        return f"Error: Bucket '{bucket_name}' not found."
    except exceptions.Forbidden as e:
        return f"Error: Permission denied for bucket '{bucket_name}'. Details: {e}"

Configuration

Hardcoding values like the default location ("US") is inflexible. These could be moved to environment variables or a configuration file for better management.

Contributing

Contributions are welcome! Please feel free to submit a pull request for bug fixes, new features, or improvements to the documentation.