Tigris

The Tigris MCP Server lets AI agents connect to Tigris buckets and objects through the MCP protocol. It enables you to manage buckets, upload or retrieve objects, create folders, generate shareable links, and perform other bucket-and-object operations from your AI workflows with secure and scalable access.

How to use

Connect your MCP client to the Tigris MCP Server to control and query your Tigris buckets and objects from AI-assisted editors and workflows. You can operate on buckets and objects using the MCP client you already use, following the patterns you typically apply for MCP integrations. Typical actions include listing buckets, creating and deleting buckets, listing objects in a bucket, uploading files, creating folders, creating or editing files with content, generating shareable links, and deleting objects.

How to install

Prerequisites vary by installation method. Choose your preferred path and follow the steps exactly as shown.

// NPX based start (local execution)
npx -y @tigrisdata/tigris-mcp-server run
// When using NPX, provide environment variables as needed in your shell or environment
// Example environment variables (placeholders):
export AWS_ACCESS_KEY_ID=YOUR_AWS_ACCESS_KEY_ID
export AWS_SECRET_ACCESS_KEY=YOUR_AWS_SECRET_ACCESS_KEY
export AWS_ENDPOINT_URL_S3=https://fly.storage.tigris.dev

Docker-based start (sandboxed container). The container runs with environment variables and a mounted path for user data. Use a path that is appropriate for your platform and user.

docker run \
  -e AWS_ACCESS_KEY_ID \
  -e AWS_SECRET_ACCESS_KEY \
  -e AWS_ENDPOINT_URL_S3 \
  -i --rm \
  --mount type=bind,src=/Users/CURRENT_USER/tigris-mcp-server,dst=/Users/CURRENT_USER/tigris-mcp-server \
  quay.io/tigrisdata/tigris-mcp-server:latest

If you prefer AWS profiles, you can enable profile usage in the same Docker or NPX flow by setting USE_AWS_PROFILES and AWS_PROFILE, along with the endpoint URL.

// NPX with AWS profiles
npx -y @tigrisdata/tigris-mcp-server run
// Environment
export USE_AWS_PROFILES=true
export AWS_PROFILE=default
export AWS_ENDPOINT_URL_S3=https://fly.storage.tigris.dev

// Docker with AWS profiles
docker run \
  -e USE_AWS_PROFILES \
  -e AWS_PROFILE \
  -e AWS_ENDPOINT_URL_S3 \
  -i --rm \
  --mount type=bind,src=/Users/CURRENT_USER/tigris-mcp-server,dst=/Users/CURRENT_USER/tigris-mcp-server \
  quay.io/tigrisdata/tigris-mcp-server:latest

Notes and configuration

The MCP server exposes actions for bucket and object management. You will typically supply your AWS-style credentials and an endpoint URL that points to the S3-compatible service. Use secure, long-lived credentials and store them in a password manager or secret store in your environment.

Environment variables shown in examples include AWS_ACCESS_KEY_ID, AWS_SECRET_ACCESS_KEY, AWS_ENDPOINT_URL_S3, USE_AWS_PROFILES, and AWS_PROFILE. These control authentication and the endpoint that MCP uses to reach the object storage service.

If you need to switch between direct credentials and profile-based authentication, choose USE_AWS_PROFILES along with AWS_PROFILE, and set AWS_ENDPOINT_URL_S3 to the appropriate endpoint.

Available tools

List Buckets

List all buckets available in your Tigris account and region.

Create Bucket

Create a new bucket with a specified name to organize objects.

Delete Bucket

Remove a bucket and all its contents.

List Objects

List objects within a specified bucket.

Upload Object

Upload a local file to a bucket, creating the object in the storage.

Create Folder

Create a logical folder within a bucket to organize objects.

Create File

Create a new file within a bucket with specified content.

Share Link

Generate a shareable link for an object.

Delete Object

Delete a specific object from a bucket.