This guide will show you how to connect Chatway MCP (Model Context Protocol) with Gemini CLI, allowing Gemini CLI to securely access your Chatway tools and data through an MCP connection.
Once connected, Gemini CLI can work with your Chatway customer support conversations and data based on the permissions you provide.
You can use Gemini CLI to find and review conversations, access relevant customer information, work with your support knowledge, and perform supported actions in Chatway.
For automatic real-time replies to incoming customer messages, connect Chatway webhooks to your AI agent and use the Chatway API to send responses. Chatway MCP can be used alongside this workflow to access Chatway tools and data.
Chatway MCP can also be used with Gemini CLIto review and summarize support information, identify recurring customer issues, and uncover useful patterns from your conversations and customer data.
Chatway’s MCP integration is available on all Chatway plans.
Requirements:
Before connecting Chatway MCP to Gemini, make sure that:
- Gemini CLI is installed
- You have a Gemini account with access to MCP integrations
- Your Chatway MCP connection details, which we’ll obtain in the steps below. You must have an Owner or Admin role in your Chatway account to authorize the MCP connection with Gemini CLI.
Obtain the Chatway MCP Connection Details
First, let’s obtain the Chatway MCP connection details you’ll need to connect Chatway with Gemini CLI. Please follow the steps below.
Step 1: Log in to your Chatway account and open the AI settings page by clicking on AI Agents or by visiting this link directly: https://go.chatway.app/ai-settings
Step 2: Go to the External AI page from the navigation on the left:

Step 3: Click Connect an AI agent:

Step 4: Select Gemini:

Step 5: Copy the Chatway MCP configuration:

Step 6: Open your terminal and add Chatway to your Gemini CLI MCP configuration using your Chatway MCP URL:
gemini mcp add --transport http chatway https://mcp.chatway.app/mcpStep 7: Verify that the Chatway MCP server was added successfully by running:
gemini mcp list
Step 8: Start Gemini CLI and authenticate the Chatway MCP connection by running the following command and click Yes to continue:
/mcp auth chatway
Step 9: Next, you will need to authorize the connection through the popup. Click on the “Connect” button:

Step 10: Choose your widget, enter a connection name and click “Continue“:

Step 11: Review the permissions that you will give to Gemini and click on the “Allow access & continue” button to allow access:

You can review and revoke the access at any time from the Chatway settings.
Step 12: Go back to your Chatway AI Agents page, and you will see the successfully connected Gemini CLI agent:

Testing your AI agent
Now let’s see how you can use the new connection.
Step 1: Open a new conversation with your Gemini CLI agent.
Step 2: Open the Gemini CLI Agent and enter your request. Gemini automatically makes connected MCP tools available in the connectors and can use them when relevant.
Here’s an example:
“Find all conversations about checkout issues from the last 7 days.”

Following this example, you may receive the answer below with all relevant data and additional details:
- ID: 1d74ad39-1e57-47c0-aab3-fa6315256e9b
- Contact: Banana #53764
- Subject: “Hi, I’m trying to place an order on your website, but I’m having troub…”
- Status: Unresolved
- Origin URL: https://www.w3schools.com/html/tryit.asp
- Assigned agent: Chatway MCP ([email protected])
- Created: 2026-09-30
- Last activity: 2026-10-21 (within the last 7 days)
- Summary of last message: The user is having trouble completing the checkout process, specifically at the payment step, and cannot finish their purchase.

For automatic real-time replies to incoming customer messages, connect Chatway webhooks to your AI agent and use the Chatway API to send responses. Chatway MCP can be used alongside this workflow to access Chatway tools and data.

That’s it! Once connected, you can use Gemini CLI to securely access Chatway conversations and contacts, search and summarize support data, and perform supported actions.