This Yandex Cloud MCP server acts as a friendly bridge between the Claude AI desktop app and Yandex Cloud’s storage system. In simple terms, it gives the AI the ability to "look" inside cloud storage folders and manage files directly through a chat interface. Instead of a user having to manually download data and upload it to the AI, they can simply ask the assistant to browse their cloud buckets or fetch specific documents, making cloud management feel like a natural conversation. On a more technical level, the tool provides three essential capabilities: listing all buckets in an account, viewing the contents of a specific bucket, and downloading files to a local directory. It uses the Model Context Protocol to expose these functions as tools that a Large Language Model can call upon autonomously. This allows the AI to perform complex operations, such as scanning large datasets for specific information or preparing local environments with cloud-hosted assets, all based on simple natural language prompts. The server is built using Node.js and TypeScript, leveraging the AWS SDK’s compatibility with Yandex Cloud’s S3-compliant endpoints. Security is a priority, as the system manages authentication through environment variables and service account keys to ensure sensitive credentials remain protected. For developers building AI-driven workflows, this MCP server is an invaluable asset because it turns static cloud storage into an active, queryable resource that an AI agent can navigate and utilize to provide deeper, data-driven insights.

1. Installation

To set up the Yandex Cloud MCP server locally, follow these steps:

Step 1: Create the project directory

mkdir yandex-cloud-mcp
cd yandex-cloud-mcp
npm init -y

Step 2: Install dependencies

# Main dependencies
npm install @modelcontextprotocol/sdk @aws-sdk/client-s3 dotenv

# Development dependencies
npm install -D typescript @types/node ts-node nodemon

Step 3: Configure TypeScript
Create a tsconfig.json file in the root directory:

{
  "compilerOptions": {
    "target": "ES2020",
    "module": "commonjs",
    "outDir": "./dist",
    "rootDir": "./src",
    "strict": true,
    "esModuleInterop": true,
    "skipLibCheck": true,
    "forceConsistentCasingInFileNames": true,
    "resolveJsonModule": true
  },
  "include": ["src/**/*"],
  "exclude": ["node_modules", "dist"]
}

Step 4: Update package.json scripts
Add the following to your package.json:

"scripts": {
  "build": "tsc",
  "start": "node dist/index.js",
  "dev": "ts-node src/index.ts",
  "watch": "nodemon --exec ts-node src/index.ts"
}

Step 5: Create source and build
1. Create a src folder.
2. Create src/index.ts and paste the server code (starting with #!/usr/bin/env node).
3. Build the project:

npm run build

---

2. Configuration

Yandex Cloud Credentials

You must obtain a Static Access Key from the Yandex Cloud Console:
1. Create a Service Account in IAM.
2. Assign roles: storage.editor (read/write) or storage.viewer (read-only).
3. Create a Static Access Key and save the Access Key ID and Secret Access Key.

Claude Desktop Configuration

Add the server to your claude_desktop_config.json file:
- Windows: %APPDATA%\Claude\claude_desktop_config.json
- macOS: ~/Library/Application Support/Claude/claude_desktop_config.json

JSON Configuration:

{
  "mcpServers": {
    "yandex-cloud": {
      "command": "node",
      "args": ["/FULL/PATH/TO/YOUR/PROJECT/yandex-cloud-mcp/dist/index.js"],
      "env": {
        "YANDEX_ACCESS_KEY_ID": "YOUR_ACCESS_KEY_ID",
        "YANDEX_SECRET_ACCESS_KEY": "YOUR_SECRET_ACCESS_KEY"
      }
    }
  }
}

Note: Replace /FULL/PATH/TO/YOUR/PROJECT/ with the actual absolute path to your project folder.

---

3. Available Tools

The server provides tools to interact with Yandex Object Storage (S3-compatible API):

Tool	Description
List Buckets	Displays a list of all buckets in your Yandex Cloud account.
View Content	Browse the files and folders within a specific bucket.
Download Files	Download files from a bucket to a local directory.

---

4. Example Prompts

Once configured, you can use prompts like these in Claude:

• "Можешь показать мои bucket в Yandex Cloud?" (Can you show my buckets in Yandex Cloud?)
• "What files are inside my [bucket-name] bucket?"
• "Download the latest log file from Yandex Cloud storage."

Use Case 1: Automated Log Analysis and Troubleshooting

Problem: DevOps engineers and developers often store application logs in Yandex Cloud Object Storage. When a production issue occurs, manually finding, downloading, and searching through these large log files to identify errors is time-consuming and tedious.
Solution: This MCP allows Claude to directly interact with your Yandex Cloud buckets. You can ask Claude to find the most recent log file, download it, and analyze the content for specific stack traces or error patterns without ever leaving the chat interface.
Example: "Claude, list the contents of the production-logs bucket. Find the log file from the last hour, download it, and tell me if there are any '500 Internal Server Error' entries."

Use Case 2: Content Audit and Storage Organization

Problem: Over time, Object Storage buckets can become cluttered with temporary files, old backups, or inconsistently named assets. Getting a high-level overview or "audit" of what is taking up space usually requires navigating through the web console or writing custom scripts.
Solution: Claude can use the MCP to list all objects in a bucket and categorize them. It can help identify large files, old versions, or files that don't follow naming conventions, providing a clear summary and suggestions for cleanup.
Example: "List all files in my marketing-assets bucket. Group them by file type and tell me which files are older than six months so I can decide what to archive."

Use Case 3: Seamless Integration for AI-Assisted Research

Problem: Researchers and data scientists often store datasets, PDFs, or research papers in cloud buckets. When they want to ask an AI questions based on those specific documents, they have to manually download the file from the cloud and then upload it back into the AI's chat window.
Solution: This MCP bridges the gap between the storage and the AI. Claude can browse the bucket, identify the relevant document the user is talking about, download it to the local environment, and then use its reasoning capabilities to answer questions or summarize the data.
Example: "Look into my project-research bucket and find the PDF about 'Q3 Market Trends'. Download it and give me a 5-bullet point summary of the main findings."

Use Case 4: Verifying CI

CD Build Artifacts
Problem: After a CI/CD pipeline runs, a developer needs to verify that the build artifacts (like a .zip or .apk file) were correctly generated and uploaded to Yandex Cloud with the correct versioning and file size.
Solution: Instead of logging into the Yandex Cloud Console, the developer can ask Claude to check the bucket. Claude can verify the existence of the file, check the timestamp, and confirm the metadata, providing a quick confirmation within the development workflow.
Example: "Check the build-releases bucket. Did the file app-v2.4.1.zip get uploaded today? If so, what is its size?"