Connect Your AI Application to Appian MCP Server

End users connect Model Context Protocol (MCP)-compatible AI applications to Appian MCP Server to interact with application data and business logic through natural language.

Appian MCP Server enforces the same object security, record-level security, and group membership as the Appian platform. Once connected, you can ask questions about your data, start processes, invoke business logic, and interact with AI agents—all from your preferred AI application.

Prerequisites for connecting to Appian MCP Server

Before you connect your AI application to Appian MCP Server via Model Context Protocol (MCP):

• Appian MCP Server enabled: A system administrator must enable Appian MCP Server in the Admin Console. If the server is not enabled, you cannot connect.
• Credentials: You need the service account API key provided by your system administrator.
• Permissions: The service account must be in assigned the Designer role. To invoke Appian AI agents through MCP, the service account must also be assigned to the Application User role.

Get your connection details

To connect your AI application to Appian MCP Server, you need the following connection details from your system administrator:

• Server URL: Your Appian environment URL followed by /mcp.
• API key: The service account API key configured for MCP access.

Your MCP configuration uses the following format:

1
2
3
4
5
6
7
8
9
10
{
  "mcpServers": {
    "appian": {
      "url": "<ENVIRONMENT_URL>/mcp",
      "headers": {
        "Authorization": "Bearer <API_KEY>"
      }
    }
  }
}

Replace <ENVIRONMENT_URL> with your Appian environment URL and <API_KEY> with the service account API key provided by your system administrator.

Configure your AI application

Add the Model Context Protocol (MCP) server configuration to your AI application's MCP settings. The following sections provide configuration instructions for specific MCP hosts.

Claude Code

To connect Claude Code to Appian MCP Server:

1. Open Claude Code.
2. Run /settings to open the MCP configuration.
3. Add the Appian MCP Server configuration to the mcpServers section of your configuration file.
4. Save the file.

Example configuration:

1
2
3
4
5
6
7
8
9
10
{
  "mcpServers": {
    "appian": {
      "url": "<ENVIRONMENT_URL>/mcp",
      "headers": {
        "Authorization": "Bearer <API_KEY>"
      }
    }
  }
}

Replace <ENVIRONMENT_URL> with your Appian environment URL and <API_KEY> with the API key provided by your system administrator.

Kiro

To connect Kiro to Appian MCP Server:

1. Open Kiro.
2. Open the MCP configuration file at .kiro/settings/mcp.json in your workspace or ~/.kiro/settings/mcp.json for global configuration.
3. Add the Appian MCP Server configuration to the mcpServers section.
4. Save the file.

Example configuration:

1
2
3
4
5
6
7
8
9
10
{
  "mcpServers": {
    "appian": {
      "url": "<ENVIRONMENT_URL>/mcp",
      "headers": {
        "Authorization": "Bearer <API_KEY>"
      }
    }
  }
}

Gemini CLI

To connect Gemini CLI to Appian MCP Server:

1. Open your Gemini CLI configuration file.
2. Add the Appian MCP Server as an MCP endpoint using the server URL and API key.
3. Save the configuration.

Authenticate

Appian MCP Server uses service account authentication. Model Context Protocol (MCP) connections authenticate using an API key provided by your system administrator.

The service account's permissions determine which tools you can discover and use. If you cannot see an expected tool, contact your system administrator to verify the service account has the appropriate permissions and that the tool has been enabled for MCP in the design object's properties.

Verify your connection

After configuring your AI application, verify the Model Context Protocol (MCP) connection is working:

1. Open your AI application.
2. Ask a question about your Appian data. For example, "What tables are available?"
3. Confirm the AI application returns results from your Appian environment.

If the connection is successful, you can see the available MCP tools from Appian in your AI application's tool list.

What you can do once connected

Once connected to Appian MCP Server via Model Context Protocol (MCP), you can use natural language to:

• Query data: Ask questions about the data in your applications, powered by your data fabric.
• Start processes: Kick off automated workflows built with process models.
• Run calculations: Perform calculations or data transformations using expression rules.
• Chat with AI agents: Send prompts to Appian AI agents and receive responses.

The AI application uses the available MCP tools to fulfill your requests. You do not need to know which specific tools to call—the AI application handles tool discovery and selection based on your natural language input.

Understand error messages

The following table describes common Model Context Protocol (MCP) error messages you may encounter:

Troubleshoot connection issues

If you cannot connect your AI application to Appian MCP Server via Model Context Protocol (MCP):

1. Verify the server URL matches the format <ENVIRONMENT_URL>/mcp.
2. Confirm the API key is correct and has not been regenerated since you received it.
3. Confirm Appian MCP Server is enabled by checking with your system administrator.
4. Restart your AI application after making configuration changes.
5. If the issue persists, contact your system administrator with the specific error message you are receiving.

Feedback