Skip to main content
Gainsight Inc.

Connect Staircase AI to LLMs Using MCP

  1. Last updated
  2. Save as PDF
This article helps admins connect their LLMs to the Staircase AI MCP server, using a connector.

This article helps admins connect their LLMs to the Staircase AI MCP server, using a connector.

Overview

Enable secure MCP access for Staircase AI users through Google or Microsoft authentication using FastMCP, allowing seamless integration with leading LLMs such as Claude, ChatGPT and Gemini across both standalone and CS/PX-integrated deployments. Once connected  to the Staircase AI MCP server, LLMs can call Staircase AI as a live data source, enabling users to receive responses enriched with real-time Staircase AI customer intelligence.

Activate Staircase AI MCP Server Access

This section explains how admins enable MCP access at the Staircase AI platform level. This is a required first step before connecting Staircase AI to any LLM.

Follow the steps below to activate Staircase AI MCP server access:

  1. Login to Staircase AI.
  2. Navigate to Settings > Configurations > Privacy & Security.
  3. Select the Enable MCP Server Access checkbox.
  4. Click Submit.

    Enable MCP server access checkbox in privacy and security

Once MCP access is enabled, you can proceed to configure individual LLM integrations.

Install Staircase AI MCP for ChatGPT

Connecting ChatGPT to the Staircase AI MCP server, ChatGPT can call Staircase AI as a live data source and users can receive answers enriched with data from Staircase AI.

Prerequisites

To configure Staircase AI MCP for ChatGPT, you must have:

  • A Staircase AI admin account
  • ChatGPT access with Developer Mode enabled
  • Admin-level access to configure apps in ChatGPT
  • An email address shared between both systems

Configure Secure MCP Access for Staircase AI Users in ChatGPT

Configuring MCP access for ChatGPT involves three steps. Perform them in the following order:

  1. Turn on Developer Mode in ChatGPT
  2. Create the Staircase AI App in ChatGPT
  3. Add the Staircase AI App in the ChatGPT

Turn on Developer Mode in ChatGPT

To activate Developer mode in ChatGpt:

  1. Log in to ChatGPT using the same email address that has access to Staircase AI.
  2. From the left navigation menu, click on your name and select Settings. The Settings dialog box appears.

    Setting option in ChatGPT profile.
     
  3. Navigate to Apps > Advanced settings. 

    Advance settings in apps option.jpg
     
  4. Turn on the Developer mode toggle.

    Developer mode toggle.

Create the Staircase AI App in ChatGPT

After Developer Mode is enabled, create the Staircase AI app:

  1. Click Create app. The New App dialog box appears. 

    New app dialog box in GPT.
     
  2. Enter the following details:
    1. Name: Staircase AI
    2. (Optional) Enter a description for the new app.
    3. MCP Server URL: https://mcp.staircase.ai/mcp
    4. Select the I understand and want to continue checkbox. 
    5. Click Create.

      I understand checkbox and create option in the new app dialog box.
       
  3. On the Cloudflare OAuth login screen, select your login method (Google or Azure).

    Cloudfare login screen with sign in options.
     
  4. Sign in using the same email address that has Staircase AI access. After successful authentication, you are redirected to ChatGPT with a confirmation message.

Add the Staircase AI App in the ChatGPT

To add Staircase AI app in ChatGPT:

  1. Open a new chat in ChatGPT. An orange border around the prompt field indicates  Developer Mode is active.
  2. In the prompt field, click the + icon.

    Developer mode in GPT.
     
  3. Hover over More and select Staircase AI from the list of apps. Once the app is loaded, the chat window is ready to send and receive enriched responses.

    GPT responds to questions asked for Staircase.

When you ask a question, ChatGPT may attempt to call the Staircase AI MCP server for additional data. A prompt appears asking you to approve each request. This approval step is a built-in security measure. A successful call shows a response from Staircase AI within the chat.

GPT pulling staircase account data.

Install Staircase AI MCP for Claude

Integrating Claude with Ask Staircase using Staircase AI MCP allows Claude to access and query customer intelligence directly from Staircase AI.

Prerequisites

To install Staircase AI MCP for Claude, you must have:

  • Access to Claude Desktop or Claude Code (CLI)
  • An active Internet connection
  • Basic familiarity with editing configuration files or running command-line interface (CLI) commands

Configure Secure MCP Access for Staircase AI Users in Claude

Follow one of the three options below to configure MCP access for Staircase AI users in Claude:

Option 1: Configure MCP for Claude Desktop

To install Staircase AI MCP for Claude Desktop:

  1. Log in to Claude. 
  2. Navigate to Settings > Developer > Edit Config.
  3. In the claude_desktop_config.json file, add the following configuration block:
    {"mcpServers": {"staircase-ai": {"command": "npx","args": ["-y", "mcp-remote", "https://mcp.staircase.ai/mcp"]}}}
  4. Restart Claude Desktop to apply the changes.

Option 2: Configure MCP for Claude using Command Line Interface (CLI)

To add Staircase MCP from the command line:
claude mcp add staircase-ai -- npx -y mcp-remote https://mcp.staircase.ai/mcp

Option 3: Configure MCP for Claude Chat (Web)

To add Staircase AI app in Claude:

  1. In the Claude prompt field, click the + icon.
  2. Click Add connectors. The Connectors dialog box appears.

    Add connectors option when you click plus.
     
  3. Click Manage connectors to view the available connectors. 

    Manage connector option in claude.
     
  4. Search for Staircase AI and click Connect.

    Connect option in Staircase connector.
     
  5. On the Cloudflare OAuth login screen, select your login method (Google or Azure).

    Cloudfare login screen with sign in options.
     
  6. Sign in using the same email address that has Staircase AI access. After successful authentication, you are redirected to the Claude Connectors page.
  7. Search for Staircase AI and click Configure. The Staircase AI connector page appears. 

    Configure option for Staircase AI connector.
     
  8. From the Tool permissions dropdown list, select Always allow.

    Tool permissions dropdown list, select Always allow..jpg
     
  9. Open a new chat in Claude and verify that Staircase AI appears in Connectors.

    Staircase AI appears in Connectors.

Verify the Integration

After completing the configuration, verify that Claude can access Staircase AI data. Ask Claude a customer-related question, such as: “What's happening with [Customer Name]?”. 

Ask questions in claude.

Include role or context cues in your questions for more relevant results. For example:

  • “I’m preparing for a call with [Customer]. What should I know?”
  • “What feedback has [Customer] shared recently?”

Install Staircase AI MCP for Gemini

You can configure Gemini Command Line Interface (CLI) to connect with Staircase AI through an MCP server. This integration allows Gemini to access customer intelligence from Staircase AI.

The Gemini web app (gemini.google.com) does not currently support custom MCP connectors. These instructions apply only to the Gemini CLI. 

Prerequisites

Before you begin, ensure the following requirements are met:

Configure Secure MCP Access for Staircase AI Users in Gemini

Follow the order below to configure MCP access for Staircase AI users in Gemini:

  1. Configure MCP for Gemini using Manual Configuration 
  2. Verify the Integration
  3. Authenticate Gemini CLI with Staircase AI
  4. Use Staircase AI in Gemini CLI

Configure MCP for Gemini using Manual Configuration

To add the MCP server using manual configuration:

  1. Open the Gemini CLI settings file:
    • Mac/Linux: ~/.gemini/settings.json
    • Windows: %USERPROFILE%\.gemini\settings.json
  2. Add the following JSON Staircase AI MCP server configuration:
    "mcpServers": {"staircase-ai": {"httpUrl": "https://mcp.staircase.ai/mcp", "oauth": {"enabled": true}}}
  3. If other MCP servers are already configured, insert the staircase-ai entry into the existing mcpServers object.
  4. Save the file.

Verify the Integration

Run this command to confirm the server is configured:  gemini mcp list
Confirm that staircase-ai appears in the list of configured servers.

Authenticate Gemini CLI with Staircase AI

To authenticate Gemini with Staircase AI:

  1. Start a new Gemini CLI session.
  2. Initiate a query that requires Staircase AI data.

    Initiate a query in Gemini.
     
  3. On the Cloudflare OAuth login screen, select your login method (Google or Azure).

    Cloudflare access page.
     
  4. Log in using a Staircase-enabled email account. After successful authentication, you are redirected back with a confirmation message.

Use Staircase AI in Gemini CLI

Once authenticated, you can query Staircase AI through Gemini CLI. For example:

  • What's happening with [Customer Name]?
  • What feature requests have customers mentioned recently?
  • What are the health signals for [Account]?

Gemini showing results upon query.

Note: You may be prompted to approve tool calls as part of Gemini’s security model.