MCP Server

Web.Casa implements a Model Context Protocol (MCP) server as a dedicated plugin, exposing 44 tools that let external AI tools and IDEs interact with your server through a standardized interface.

What Is MCP

MCP (Model Context Protocol) is a standard that exposes server management capabilities as structured tools that AI models can call. By connecting an MCP-compatible client to Web.Casa’s MCP server, your AI assistant gains the ability to manage hosts, containers, files, databases, and more — all through natural language.

API Token Management

Access to the MCP server is authenticated via API tokens with scope-based permissions.

Creating a Token

1. Navigate to Settings > MCP Server
2. Click Create Token
3. Choose an expiry option:

Expiry	Use Case
30 days	Short-term testing or temporary access
90 days	Regular development workflows
180 days	Long-running integrations
365 days	Annual renewal cycle
No expiry	Persistent integrations (use with caution)

4. Copy the full token immediately — it is only shown once

Token Security

• Tokens are prefixed with wc_ for easy identification
• Stored as SHA-256 hashes — the raw token is shown only once at creation
• Each token tracks creation date, expiry, and last used timestamp
• Permission scopes use category:action format (e.g., hosts:write, deploy:read)
• Wildcard *:* grants full access
• Tokens can be revoked immediately from the settings UI

Tip: Rotate tokens periodically. If a token may have been compromised, revoke it immediately and create a new one.

MCP Endpoint

https://your-server.com/api/plugins/mcpserver/mcp

IDE Configuration

Cursor

{
  "mcpServers": {
    "webcasa": {
      "url": "https://your-server.com/api/plugins/mcpserver/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_TOKEN"
      }
    }
  }
}

Windsurf

{
  "mcpServers": {
    "webcasa": {
      "url": "https://your-server.com/api/plugins/mcpserver/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_TOKEN"
      }
    }
  }
}

Claude Code

{
  "mcpServers": {
    "webcasa": {
      "url": "https://your-server.com/api/plugins/mcpserver/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_TOKEN"
      }
    }
  }
}

Tip: The configuration snippets on the MCP Server settings page are pre-filled with your actual server URL. Copy them directly and only replace the token value.

Available Tools (44 Total)

Host Management

• list_hosts — list all reverse proxy hosts
• create_host — create a new host with domain, upstream, and TLS settings
• update_host — update upstream, TLS mode, WebSocket, compression, or enabled state
• delete_host — remove a host by ID
• toggle_host — enable or disable a host

Project Deployment

• list_projects — list all deployment projects
• get_project — get project details
• deploy_project — create a project from Git and trigger the first build
• build_project — trigger a new build for an existing project
• get_build_logs — get build logs for a project
• auto_deploy — one-command: Git URL → project → build → reverse proxy with HTTPS

Docker

• docker_ps — list all containers with status, ports, and resource usage
• docker_logs — get container logs
• docker_run — run a new container with image, ports, volumes, and env vars
• list_stacks — list Docker Compose stacks
• create_stack — create a Compose stack from YAML
• control_stack — up, down, or restart a Compose stack
• get_stack_logs — get stack logs

Database

• list_db_instances — list all managed database instances
• create_db_instance — create a new database instance
• execute_query — execute a read-only SQL query

System & Files

• get_system_metrics — CPU, memory, disk usage, load average
• get_system_info — hostname, OS, kernel, uptime, CPU cores, architecture
• run_command — execute a shell command with timeout
• read_file — read file contents
• list_directory — list files and directories
• write_file — write content to a file
• delete_file — delete a file

AI-Powered

• generate_compose — generate Docker Compose YAML from natural language
• diagnose_error — analyze error logs with AI for root causes and fixes

Monitoring & Alerts

• list_alert_rules — list all alert rules
• create_alert_rule — create a monitoring alert rule
• get_alert_history — get recent alerts

Firewall

• firewall_status — get firewall status and active zones
• manage_firewall_rule — add or remove port or service rules

Notifications

• list_notify_channels — list configured notification channels

Backup

• trigger_backup — trigger an immediate system backup

System Inspection

• run_inspection — run a health inspection with AI-powered analysis

Cron Jobs

• list_cron_jobs — list all scheduled jobs (optional tag filter)
• get_cron_job_logs — get execution logs
• create_cron_job — create a scheduled job
• update_cron_job — update job settings
• delete_cron_job — delete a job
• trigger_cron_job — manually trigger a job

Security Considerations

• Always use HTTPS for the MCP endpoint to protect tokens in transit
• Create separate tokens for different clients or team members
• Use scoped permissions — grant only the categories each client needs
• Monitor the last used timestamp to detect unused or potentially compromised tokens
• Revoke tokens when they are no longer needed