# Cloudru S3 MCP Ai Cli

> Cloudru S3 MCP Ai Cli is an open-source MCP server for cloud and infrastructure. Go MCP server and CLI for Cloud.ru Evolution Object Storage, read-only by default. It runs locally on your machine and is started by your AI client.

- **Kind:** MCP server
- **Pricing:** Open source
- **Platforms:** CLI, self-hosted, Linux, macOS, Windows
- **Repository:** https://github.com/stufently/cloudru-s3-mcp-ai-cli
- **MCP Registry name:** io.github.stufently/cloudru-s3-mcp-ai-cli
- **Categories:** [cloud and infrastructure](https://ai-powered.com/categories/cloud-and-infrastructure), [developer tools](https://ai-powered.com/categories/developer-tools)
- **Listing updated:** 2026-11-18

## About

The binary is called `cloudru-s3` and is written for Cloud.ru's Evolution Object Storage, with the `s3.cloud.ru` endpoint and the `ru-central-1` signing pool. The README is explicit that it is not meant for the other Cloud.ru S3 product, which uses a different endpoint and region and needs its own profile.

Access starts at read level. Write, delete and admin tools are switched on one by one with `--allow-write`, `--allow-delete`, `--allow-admin` or the `CLOUDRU_S3_ALLOW` variable, and destructive actions follow a policy that asks for confirmation by default. Bucket and prefix allow-lists and an optional confined local file root keep an agent inside the area you choose, and an append-only audit file can record what it did.

The same binary doubles as a command-line client (`ls`, `cp`, `du`, `presign`, `doctor`) and installs its own MCP config for Cursor, Claude Desktop, Codex and Claude Code. It ships as release binaries or a container image run as a non-root user. The authors note that the profile and the cloud API still need live verification.

## Features

- **Evolution profile:** Targets s3.cloud.ru in the ru-central-1 pool; a different Cloud.ru product needs a separate profile.
- **Opt-in permission levels:** Read is the default; write, delete and admin tools are enabled independently.
- **Bucket and prefix scope:** Comma-separated allow-lists restrict which buckets and prefixes an agent can touch.
- **Confirmation for destructive calls:** The destructive policy can confirm, elicit, block or allow, and defaults to confirm.
- **Bounded scans:** Size and prefix-usage analytics stop at a scan limit (10000 by default) and flag partial results.
- **Presigned and multipart transfers:** Large files go through presigned URLs, multipart uploads or a confined file root.
- **Built-in CLI and installer:** The same binary lists, copies and diagnoses, and writes MCP config for several clients.

## Good for

- Check access to an Evolution bucket before a migration
- Restore a backup archive from a bucket to a local file
- Find the largest objects or unfinished uploads under a prefix
- Give a coding agent read-only access to a single bucket
- Create short-lived download links for existing files

## Install

### Claude Code (Terminal)

```sh
claude mcp add cloudru-s3-mcp-ai-cli --env CLOUDRU_S3_ACCESS_KEY_ID='<CLOUDRU_S3_ACCESS_KEY_ID>' --env CLOUDRU_S3_SECRET_ACCESS_KEY='<CLOUDRU_S3_SECRET_ACCESS_KEY>' -- docker run -i --rm -e CLOUDRU_S3_ACCESS_KEY_ID -e CLOUDRU_S3_SECRET_ACCESS_KEY ghcr.io/stufently/cloudru-s3-mcp-ai-cli:0.1.1
```

### Claude Desktop (claude_desktop_config.json)

```json
{
  "mcpServers": {
    "cloudru-s3-mcp-ai-cli": {
      "command": "docker",
      "args": [
        "run",
        "-i",
        "--rm",
        "-e",
        "CLOUDRU_S3_ACCESS_KEY_ID",
        "-e",
        "CLOUDRU_S3_SECRET_ACCESS_KEY",
        "ghcr.io/stufently/cloudru-s3-mcp-ai-cli:0.1.1"
      ],
      "env": {
        "CLOUDRU_S3_ACCESS_KEY_ID": "<CLOUDRU_S3_ACCESS_KEY_ID>",
        "CLOUDRU_S3_SECRET_ACCESS_KEY": "<CLOUDRU_S3_SECRET_ACCESS_KEY>"
      }
    }
  }
}
```

### Cursor (~/.cursor/mcp.json)

```json
{
  "mcpServers": {
    "cloudru-s3-mcp-ai-cli": {
      "command": "docker",
      "args": [
        "run",
        "-i",
        "--rm",
        "-e",
        "CLOUDRU_S3_ACCESS_KEY_ID",
        "-e",
        "CLOUDRU_S3_SECRET_ACCESS_KEY",
        "ghcr.io/stufently/cloudru-s3-mcp-ai-cli:0.1.1"
      ],
      "env": {
        "CLOUDRU_S3_ACCESS_KEY_ID": "<CLOUDRU_S3_ACCESS_KEY_ID>",
        "CLOUDRU_S3_SECRET_ACCESS_KEY": "<CLOUDRU_S3_SECRET_ACCESS_KEY>"
      }
    }
  }
}
```

### VS Code (.vscode/mcp.json)

```json
{
  "servers": {
    "cloudru-s3-mcp-ai-cli": {
      "type": "stdio",
      "command": "docker",
      "args": [
        "run",
        "-i",
        "--rm",
        "-e",
        "CLOUDRU_S3_ACCESS_KEY_ID",
        "-e",
        "CLOUDRU_S3_SECRET_ACCESS_KEY",
        "ghcr.io/stufently/cloudru-s3-mcp-ai-cli:0.1.1"
      ],
      "env": {
        "CLOUDRU_S3_ACCESS_KEY_ID": "<CLOUDRU_S3_ACCESS_KEY_ID>",
        "CLOUDRU_S3_SECRET_ACCESS_KEY": "<CLOUDRU_S3_SECRET_ACCESS_KEY>"
      }
    }
  }
}
```

### Windsurf (~/.codeium/windsurf/mcp_config.json)

```json
{
  "mcpServers": {
    "cloudru-s3-mcp-ai-cli": {
      "command": "docker",
      "args": [
        "run",
        "-i",
        "--rm",
        "-e",
        "CLOUDRU_S3_ACCESS_KEY_ID",
        "-e",
        "CLOUDRU_S3_SECRET_ACCESS_KEY",
        "ghcr.io/stufently/cloudru-s3-mcp-ai-cli:0.1.1"
      ],
      "env": {
        "CLOUDRU_S3_ACCESS_KEY_ID": "<CLOUDRU_S3_ACCESS_KEY_ID>",
        "CLOUDRU_S3_SECRET_ACCESS_KEY": "<CLOUDRU_S3_SECRET_ACCESS_KEY>"
      }
    }
  }
}
```

### Other clients (your client's MCP settings (most use this shape))

```json
{
  "mcpServers": {
    "cloudru-s3-mcp-ai-cli": {
      "command": "docker",
      "args": [
        "run",
        "-i",
        "--rm",
        "-e",
        "CLOUDRU_S3_ACCESS_KEY_ID",
        "-e",
        "CLOUDRU_S3_SECRET_ACCESS_KEY",
        "ghcr.io/stufently/cloudru-s3-mcp-ai-cli:0.1.1"
      ],
      "env": {
        "CLOUDRU_S3_ACCESS_KEY_ID": "<CLOUDRU_S3_ACCESS_KEY_ID>",
        "CLOUDRU_S3_SECRET_ACCESS_KEY": "<CLOUDRU_S3_SECRET_ACCESS_KEY>"
      }
    }
  }
}
```

## Tools

- `s3_list_buckets`: Lists the buckets the credentials can see.
- `s3_list_objects`: Lists objects under a prefix, up to 1000 per page.
- `s3_head_object`: Returns an object's metadata.
- `s3_get_object`: Reads an object inline, up to 262144 bytes.
- `s3_preview_object`: Returns a short preview of an object, up to 65536 bytes.
- `s3_presign_get`: Creates a presigned download URL.
- `s3_list_object_versions`: Lists versions of objects.
- `s3_list_multipart_uploads`: Lists multipart uploads in progress.
- `s3_list_parts`: Lists the parts of a multipart upload.
- `s3_get_bucket_config`: Reads a bucket's configuration.
- `s3_largest_objects`: Finds the largest objects under a prefix within a bounded scan.
- `s3_prefix_usage`: Reports aggregated usage per prefix in pages.
- `s3_unfinished_uploads`: Finds abandoned multipart uploads.
- `s3_download_file`: Downloads an object to the confined local file root.
- `s3_get_object_tagging`: Reads an object's tags.
- `s3_put_object`: Writes an object inline.
- `s3_create_folder`: Creates a folder marker.
- `s3_copy_object`: Copies an object.
- `s3_presign_put`: Creates a presigned upload URL.
- `s3_multipart_create`: Starts a multipart upload.
- `s3_multipart_presign_part`: Presigns one part of a multipart upload.
- `s3_multipart_complete`: Completes a multipart upload.
- `s3_upload_file`: Uploads a file from the confined local file root.
- `s3_put_object_tagging`: Writes an object's tags.
- `s3_delete_object`: Deletes one object.
- `s3_delete_objects`: Deletes several objects.
- `s3_multipart_abort`: Aborts a multipart upload.
- `s3_create_bucket`: Creates a bucket.
- `s3_put_bucket_versioning`: Sets bucket versioning.
- `s3_put_bucket_cors`: Sets bucket CORS rules.
- `s3_put_bucket_policy`: Sets a bucket policy.
- `s3_delete_bucket_policy`: Removes a bucket policy.
- `s3_put_object_lock_configuration`: Sets the Object Lock configuration.
- `s3_delete_bucket`: Deletes a bucket.

## FAQ

### Which Cloud.ru storage does it support?

The Evolution Object Storage profile at s3.cloud.ru in the ru-central-1 pool. The Advanced profile needs a separate setup.

### Can the agent delete files by default?

No. Only read tools are active until you enable write, delete or admin levels, and destructive calls default to a confirmation policy.

### How do I run it?

Use a release binary, build it with Go, or run the ghcr.io container image with your access key and secret passed as environment variables.

### Where are credentials stored?

Entered through `auth set`, they go to a credentials file created with 0600 permissions; flags and environment variables take priority over it.

### Is Cloudru S3 MCP Ai Cli free?

Yes. Cloudru S3 MCP Ai Cli is open source, so you can use and self-host it for free.

### How do I add Cloudru S3 MCP Ai Cli to Claude Code?

Run this command in your terminal: claude mcp add cloudru-s3-mcp-ai-cli --env CLOUDRU_S3_ACCESS_KEY_ID='<CLOUDRU_S3_ACCESS_KEY_ID>' --env CLOUDRU_S3_SECRET_ACCESS_KEY='<CLOUDRU_S3_SECRET_ACCESS_KEY>' -- docker run -i --rm -e CLOUDRU_S3_ACCESS_KEY_ID -e CLOUDRU_S3_SECRET_ACCESS_KEY ghcr.io/stufently/cloudru-s3-mcp-ai-cli:0.1.1

### What are the best alternatives to Cloudru S3 MCP Ai Cli?

The closest alternatives to Cloudru S3 MCP Ai Cli are REA, Context7, Chrome DevTools MCP, Codebase Memory and GitHub.

## Where this comes from

- Description: written from the project's README.
- Install steps: from the official MCP Registry entry.

## Alternatives

- [REA](https://ai-powered.com/mcp/rea): REA is an open-source MCP server for developer tools.
- [Context7](https://ai-powered.com/mcp/context7): Context7 is a freemium MCP server for developer tools.
- [Chrome DevTools MCP](https://ai-powered.com/mcp/chrome-devtools-mcp-2): Chrome DevTools MCP is an open-source MCP server for browser automation.
- [Codebase Memory](https://ai-powered.com/mcp/codebase-memory): Codebase Memory is an open-source MCP server for knowledge and memory.
- [GitHub](https://ai-powered.com/mcp/github-2): GitHub is an open-source MCP server for developer tools.
- [Serena MCP: the IDE for your agent](https://ai-powered.com/mcp/serena-mcp-the-ide-for-your-agent): Serena MCP: the IDE for your agent is an open-source MCP server for developer tools.
- [awesome-mcp-servers](https://ai-powered.com/mcp/awesome-mcp-servers-2): awesome-mcp-servers is an open-source MCP server for developer tools.
- [XERJ](https://ai-powered.com/mcp/xerj): XERJ is an open-source MCP server for knowledge and memory.

---
Source: https://ai-powered.com/mcp/cloudru-s3-mcp-ai-cli
