cloudconnexa

This MCP server exposes an interface to manage and query your CloudConnexa OpenVPN environment through standardized MCP tools. After running the container, you can connect any MCP client (for example 5ire, Claude Desktop, Cursor, or VS Code with the appropriate plugin) and discover the available tools. The server organizes capabilities into categories such as Hosts, Networks, Traffic and Sessions, and Users/Devices. Use the MCP client to request tool discovery with a prompt like: Discover graphql_schema and tell me what tools are available?. The client will present a list of tools you can invoke, such as hosts, networks, networkRoutes, visitedDomains, sessions, users, and more. Each tool accepts parameters as described in the tools table, enabling read-only queries against your CloudConnexa environment. You can compose prompts to fetch hosts, networks, applications, routes, or user information and have the AI assistant execute these actions for you.

To get started in 5ire App, configure a local tool with the URL http://0.0.0.0:9999/sse, enable the MCP server, and provide your LLM provider API key. From there you can open a chat, select an OpenVPN MCP model (or your preferred LLM), and prompt the assistant to discover tools and perform actions such as listing hosts, retrieving network routes, or showing DNS logs for a timeframe. The available tool descriptions in the README will guide you on the exact syntax for each call.

Example workflows include: listing all hosts and their IPs, listing networks and their IDs, inspecting network services for a given network, and querying session records over a date range. As you experiment, you can refine prompts to filter results (e.g., by hostId, networkId, or time window) and to present results in a concise, readable format.

Prerequisites:

• Docker installed and running on your machine
• Access to a terminal or command prompt
• (Optional) Docker Compose if you prefer orchestration

Step-by-step installation:

1. Clone or download the MCP server repository to your machine.
2. Build the Docker image for the MCP server (as suggested in the README): docker build -t mcp-server-public .
3. Run the MCP server container: docker run -i -p 9999:9999 mcp-server-public This starts the MCP server and exposes the SSE endpoint at http://<host>:9999/sse
4. Verify the server is running by visiting the SSE URL or checking container logs: docker ps docker logs <container_id>
5. In your MCP client (e.g., 5ire App), add a new tool with URL http://0.0.0.0:9999/sse and ensure the server is ON.

Prerequisites notes:

• Ensure the host firewall allows inbound traffic on port 9999 if you need remote access.
• If you prefer a non-Docker setup, you can adapt the commands to run the server via the repository’s start script as described in the README.

Tips and common issues:

• If the MCP client cannot discover tools, first ensure the server is reachable at http://0.0.0.0:9999/sse from the client network. Check Docker port mappings and container status.
• The tools provided are read-only by default; prompts can be crafted to request specific information such as host lists, network routes, or session data.
• When adding GPT-5 or other models in clients like 5ire, follow the model configuration steps in the README to enable Tools and set appropriate pricing/context if needed.
• If you encounter SSL or CORS-related prompts, ensure the MCP server is accessed over a compatible endpoint and that the client is configured to trust the server.