Apono Assist MCP Server

Initiate and review just-in-time access requests inside your AI tools

A Model Context Protocol (MCP) server lets AI clients connect to external tools and data sources in a consistent, secure way.

The Apono Assist MCP Server brings Apono’s just-in-time (JIT) access capabilities into your clients:

• Efficiency: Request access without leaving your workflow

• Security and compliance: Access is controlled, logged, and tied to your Apono policies.

• Consistency: The same access workflows are available across all supported AI tools

---

Install the MCP server

Apono supports two methods for connecting your AI client to the MCP server.

Implementation type	Description
stdio (local)	Recommended option for most desktop clients In this mode, the client launches the Apono CLI to start the MCP server on your machine.
http (remote)	Required for web-based clients or for situations where the Apono CLI cannot be run locally In this mode, the client connects directly to the hosted Apono MCP endpoint.

Use stdio when possible. Choose http only when your client runs in a browser or when you cannot run the CLI locally.

Install on Cursor

stdio

Prerequisite:

The Apono CLI is required to run the MCP Server and authenticate your AI client with Apono.

• Minimum required version: 1.3.1

• Learn how to install or update the Apono CLI.

Follow these steps to install the MCP server using stdio:

1. Ensure Apono CLI is running in minimal version 1.3.1 before authenticating and installing the MCP server:

   1. From your terminal, run apono version

2. From your terminal, run apono login to authenticate your MCP client with Apono.

apono login

1. Click the following button to install the MCP Server. The MCP page in Cursor opens.

Add to Cursor

1. Click Install.

2. Quit and restart Cursor.

http

Follow these steps to install the MCP server using http:

1. Generate a Personal API Token by following this guide.

Before end users can generate their own API tokens, you need to explicitly enable this feature for your organization. To enable API token generation, please reach out to Apono Support at [email protected].

1. Click the following button to install the MCP Server. The MCP page in Cursor opens.

Add to Cursor

1. Insert the Personal API Token obtained in Step 1 into the Headers section.

2. Click Install.

3. Find the newly installed server and ensure the toggle is switched on.

4. Quit and restart Cursor.

---

Install on GitHub Copilot (relevant for most IDEs)

Prerequisite:

The Apono CLI is required to run the MCP Server and authenticate your AI client with Apono.

• Minimum required version: 1.3.1

• Learn how to install or update the Apono CLI.

Follow these steps to install the MCP server using the stdio method:

1. Ensure Apono CLI is running in minimal version 1.3.1 before authenticating and installing the MCP server:

   1. From your terminal, run apono version

2. From your terminal, run apono login to authenticate your MCP client with Apono.

apono login

1. In your IDE, open the Github Copilot settings

2. Navigate to MCP Settings > Configure. The mcp.json file opens in a new tab.

3. Add the following configuration to the file.

stdio

{
  "mcpServers": {
    "apono-access-assistant": {
      "command": "apono",
      "args": ["mcp"]
    }
  }
}

1. Save the file.

2. Make sure Copilot is running in Agent mode (the default is Ask mode). See how to change modes.

3. Quit and restart your IDE.

Read Extending GitHub Copilot Chat with the Model Context Protocol (MCP) for more information.

---

Install on Claude

Prerequisite:

1. Apono CLI: The Apono CLI is required to run the MCP Server and authenticate your AI client with Apono.

   1. Minimum required version: 1.3.1

   2. Learn how to install or update the Apono CLI.

2. Claude Desktop App: MCP integrations are currently supported only within the Claude desktop environment.

Follow these steps to install the MCP server using the stdio method:

1. Ensure Apono CLI is running in minimal version 1.3.1 before authenticating and installing the MCP server:

   1. From your terminal, run apono version

2. From your terminal, run apono login to authenticate your MCP client with Apono.

apono login

1. In Claude, from the main navigation, click Claude > Settings > Developer. The Settings page appears.

2. Click Edit Config. The claude_desktop_config.json file opens.

3. Add the stdio configuration to the file.

stdio

{
  "mcpServers": {
    "apono-access-assistant": {
      "command": "apono",
      "args": ["mcp"]
    }
  }
}

1. Save the file.

2. Quit and restart Claude desktop.

---

Use the MCP server

Follow these steps to make inquiries to Apono:

1. In your AI client, open the chat or inline interface.

2. Start a conversation in your AI client to request new access, check the status of an existing request, or retrieve access details for a resource.

Sample prompts
Can you request access for me to [resource] for 1 hour? What is the status of [request_id]? List my available resources

After submitting a request, you will receive a notification through Slack, Backstage, Teams, or email.

If issues occur while using the MCP server, troubleshoot the server.

---

Security notes

To use Apono safely, keep in mind how authorization and authentication work in your AI client.

Authorization

When you first run an Apono tool, your AI client may prompt you to authorize it. Select Allow once to ensure you maintain control over when and how each tool can be used.

Authentication

Your AI client connects to Apono through the Apono CLI. Authenticate by running apono login in your terminal. This will prompt a browser tab to open, authenticating you via SSO. You may need to re-authenticate periodically.

---

Available tools

When you use Apono through your AI client, the client communicates with the MCP server by calling specific Apono commands. Each command maps to a tool that handles a different part of the access request workflow. Unlike most MCP Servers that are based on disparate tools, the Apono Assist MCP is based on an AI Agent trained specifically to help you get access, with a "DevOps brain".

The table below lists the available commands and their functions.

Command	Description
user_get_access_details	Retrieves details for an access request, including a secure Apono Portal link. IMPORTANT: Apono never exposes credentials to the AI tool. Instead, it gives commands or secure links, or provides a link to the Apono Web Portal to consume credentials there securely.
user_get_request_status	Checks the current status of an access request
user_create_request	Submits a just-in-time access request for a resource, integration, and permission For security reasons, the Assistant only has visibility into your own requests. You cannot query or duplicate access requests made by other users. IMPORTANT: The Apono Assistant does not currently support bundles.
user_apono_assist	Explores integrations, resources, and permissions available to the user to guide and prepare for creating a request

---

Troubleshoot the server

If the MCP server does not appear or tools are unresponsive (for example, you see no prompts or tools detected), try these steps in order:

1. Check the Apono CLI version. If the version is less than 1.3.0, upgrade the CLI installation.

apono version

1. Reset your authentication.

apono logout
apono login

1. Quit and reopen your AI client. MCP servers load when the client starts.

2. Verify the MCP installation.

3. If you still don't see the MCP tools, uninstall the MCP server and reinstall it after you've verified the Apono CLI version and logged in.

4. Enable debug logging.

apono mcp --debug

To enable persistent logging, add "--debug" to the args in your configuration file.

"args": ["mcp", "--debug"]

By default, logs are written to platform-specific locations.

MacOS

$HOME/Library/Application\ Support/apono-cli/mcp_logging.log

Windows

%APPDATA%\apono-cli\mcp_logging.log

Linux

$XDG_CONFIG_HOME/apono-cli/mcp_logging.log

1. If the problem persists, send the mcp_logging.log file to Apono Support.