Remote MCP Server: Self-Hosted Deployment (Beta)

The Remote MCP setup lets your organization host a single, centrally managed MCP Server instance and share it with all users over HTTPS - eliminating the need for per-user Docker installs, local credential management, or manual update cycles.

With the Docker and NPM setups, every user in your organization installs and runs their own local MCP Server process. Instead the Remote MCP setup deploys a single shared instance on a host machine. Authentication, updates, and credential management are all handled centrally - end users just point their MCP client at a URL and connect.

Getting Started

1. Register an OAuth Client
2. Configuring Environment Variables
3. Run the Remote MCP Container
4. Configure MCP Clients

Step 1: Register an OAuth Client

Unlike the local setups (Docker and NPM) which use a Self-Client application, the Remote MCP Server requires a server-based OAuth application that supports Redirect URIs. This is because the server manages the full OAuth authorization flow on behalf of users connecting via the browser.

1. Go to the Zoho Developer Console.
2. Click Add Client and select Server-based Application (not Self-Client) as the client type.

3. Set the Authorized Redirect URI to:
   Redirect URI: 

    <your-public-url>/auth/callback

   Copy

   Replace <your-public-url> with the public HTTPS URL where you plan to host the Remote MCP Server.

4. Save the application and note down the Client ID and Client Secret. These will be used in the environment variables in the next step.

Step 2: Configuring Environment Variables

The Remote MCP Server is configured entirely through a .env file on the host machine. This file is passed to the Docker container at runtime using the --env-file flag. Create this file on your host before proceeding to the next step.

Required Environment Variables

Configure these essential variables before integrating with the MCP Hosts:

Step 3: Run the Remote MCP Container

The Remote MCP Server runs from the same Docker image as the local setup but uses the remote-v1 tag, which includes the HTTP server and authentication layer required for shared, multi-user operation. Start the container on your host machine using the command below.

 docker run --name remote-mcp-server \

  --network=host \

  --env-file .env \

zohoanalytics/mcp-server:remote-v1

Step 4: Configure MCP Clients

Once the Remote MCP Server is running and reachable at its public URL, users do not need to install or configure anything locally. All they need to do is add a short entry to their MCP host configuration file pointing at the hosted server URL.

Any MCP host that supports the HTTP transport can connect to your Remote MCP Server. The configuration is the same regardless of which AI tool the user is connecting from - only the file location and enclosing key differ per host.

Client Configuration

Add the following to the MCP host configuration file. Replace <hosted-mcp-server-url> with the actual MCP_SERVER_PUBLIC_URL value from your deployment.

{

 "mcpServers": {

   "ZohoAnalyticsMCP": {

     "url":  "<hosted-mcp-server-url>/mcp",

     "type": "http"

   }

 }

}

Tools in Zoho Analytics MCP Server

The Zoho Analytics MCP server provides various tools for interacting with Zoho Analytics.

Help & Support

We offer 24x5 technical support (Monday to Friday). In case if you need our assistance kindly do mail us your questions to support@zohoanalytics.com.

You can also reach out to us on our toll-free numbers.
United States: +1 (888) 900 9646
United Kingdom: +44 (20) 35647890
Australia: +61-2-80662898
India: +91 44 6965 6060