概览
What is Tigris MCP Server?
The Tigris MCP Server implements the Model Context Protocol (MCP) to connect AI agents with Tigris S3-compatible object storage. It enables agents to perform bucket and object management tasks directly from AI editor workflows.
How to use Tigris MCP Server?
Sign up for Tigris at storage.new and obtain an AWS_ACCESS_KEY_ID and AWS_SECRET_ACCESS_KEY. Install the server via npx (@tigrisdata/tigris-mcp-server) or Docker, and configure the required environment variables (AWS_ACCESS_KEY_ID, AWS_SECRET_ACCESS_KEY, AWS_ENDPOINT_URL_S3). For one-click setup in VS Code or VS Code Insiders, use the provided install buttons. For Claude Desktop or Cursor, run the init script npx -y @tigrisdata/tigris-mcp-server init or manually add the configuration block. Docker usage is recommended for better sandboxing and restricts file operations to a specific directory.
Key features of Tigris MCP Server
- List, create, and delete Tigris buckets.
- List objects within a bucket.
- Upload local files to a bucket.
- Create folders and files with content in a bucket.
- Generate shareable links for objects.
- Delete files from a bucket.
Use cases of Tigris MCP Server
- Manage Tigris storage buckets directly from an AI coding assistant.
- Upload and organize AI training data or models without leaving the editor.
- Share files with collaborators by generating temporary links.
- Automate backup routines by creating and deleting buckets via agent commands.
FAQ from Tigris MCP Server
Does Tigris MCP Server support OAuth?
Yes, Tigris has a hosted MCP server with OAuth support. Read more at mcp.storage.dev.
What are the runtime requirements?
The server can run via npx (requires Node.js) or Docker (requires Docker Engine). Docker is recommended for enhanced sandboxing.
How does the Docker sandbox work?
When running with Docker, the server only allows file operations within the /Users/CURRENT_USER/tigris-mcp-server directory, providing a secure environment.
Can I use existing AWS profiles instead of access keys?
Yes, set USE_AWS_PROFILES to true and specify the AWS_PROFILE name in the environment configuration.
Where do data and credentials live?
Credentials are provided via environment variables (or AWS profiles). Data is stored in your Tigris buckets, which are S3-compatible and managed through the server.
