Skip to main content
Glama

Rqeim Cloud MCP

Connect an AI agent to Rqeim Cloud and it can create, deploy, inspect and diagnose the applications on your account: environments, previews, logs, domains, health and backups.

This repository is the public distribution bundle for the Rqeim Cloud MCP server. It contains no server code: the server is hosted at https://cloud.rqeim.com/mcp.


1. What the server does

Rqeim Cloud runs Docker applications from a Git repository: you push, it deploys. With this server an agent works the same way a person does in the dashboard:

  • read the account's projects, environments, health and deploy history;

  • open a copy of a project, change its code in a sandboxed machine, preview it, and publish it;

  • read logs, run scheduled jobs, restore a backup (with explicit confirmation).

Every change is asked for explicitly by the person in the conversation. The server never decides on its own to deploy, delete, or touch production.

Related MCP server: Somewhere MCP

2. Connect

Claude Code

claude mcp add --transport http rqeim-cloud https://cloud.rqeim.com/mcp

Or install the plugin, which also adds the workflow skill:

claude plugin marketplace add Rqeim/rqeim-cloud-mcp
claude plugin install rqeim-cloud@rqeim

Claude Code opens a browser to sign in to your Rqeim account the first time.

Gemini CLI

gemini extensions install https://github.com/Rqeim/rqeim-cloud-mcp

ChatGPT and Claude (web, desktop, mobile)

Add a connector with the URL https://cloud.rqeim.com/mcp/directory and sign in when asked. /mcp/directory lists every tool by name, which is what a public directory reviews; the plain /mcp keeps the short list and its search for agents. Each door has its own OAuth resource.

Codex and other MCP clients

Streamable HTTP at https://cloud.rqeim.com/mcp, OAuth 2.1 with dynamic client registration and PKCE. An account API token works too, as Authorization: Bearer <token>.

3. Authentication

The server signs users in with OAuth 2.1, discovered from the standard well-known documents:

Document

URL

Protected resource metadata (RFC 9728)

https://cloud.rqeim.com/.well-known/oauth-protected-resource/mcp

Authorization server metadata (RFC 8414)

https://cloud.rqeim.com/.well-known/oauth-authorization-server

Clients register themselves (/oauth/register), use PKCE with S256, and receive a token scoped to mcp alone. A token is an hour long, refreshed without a new sign-in.

An unauthenticated request answers:

HTTP/1.1 401 Unauthorized
WWW-Authenticate: Bearer realm="rqeim-cloud", resource_metadata="https://cloud.rqeim.com/.well-known/oauth-protected-resource/mcp", scope="mcp"

4. Tools

Every tool is one intent. Reads change nothing and answer without confirmation; tools that change something take a requestId (generated when omitted) and destructive or production tools also need confirm: true after the person agrees. A change runs in the background and returns a task key: follow it with get_task.

Naming: list_projects, get_project, get_projects_health, create_environment, workspace_shell, read_environment_logs, finish_task, delete_environment, get_task, read_guide.

Run read_guide once per conversation: it is the current, complete description of the workflow and the rules (written for the model, kept up to date by the service).

5. Data and privacy

The server reads and writes only the account that signed in. It stores no conversation content and returns no telemetry identifiers in tool results. See PRIVACY.md.

6. Limits

  • One project belongs to one account; there are no shared workspaces.

  • Machine operations (deploy, build, restore) are asynchronous: a redeploy is 10–40 seconds, a first deploy about a minute.

  • Production is read-and-diagnose for agents: a change goes through an environment first, then a deploy.

7. Support

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    Related MCP Connectors

    Related MCP Servers

    • A
      license
      Not graded
      quality
      B
      maintenance
      Enables AI assistants to build, deploy, and inspect apps through a hosted backend service, including reading platform docs, deploying raw source, inspecting logs, verifying live pages in browsers, and managing project tasks.
      MIT
    • A
      license
      A
      quality
      B
      maintenance
      Enables AI agents to manage self-hosted Coolify v4 instances, including listing applications, safely handling environment variables with masking and dry runs, deploying, restarting, and reading logs.
      20
      MIT