• Prerequisites
• Connect through the Gemini Command Line Interface
• Connect through the Google ADK

Gemini is Google’s family of large language models (LLMs). It can generate text, images, audio, video, and code. This page explains how to connect the Connect AI MCP to Gemini through either the command-line interface or the Google Agent Development Kit (ADK).

Prerequisites

Before you can configure and use Gemini with Connect AI, you must first connect a data source to your Connect AI account. See Sources for more information.

You must also generate a Personal Access Token (PAT) on the Settings page. Copy this down, as it acts as your password during authentication. You must also generate an OAuth JWT bearer token. Copy this down, as it acts as your password during authentication.

You need Python >= 3.10 to use the Gemini tools.

Connect through the Gemini Command Line Interface

After you install Gemini, do the following:

1. Go to your user directory %USERPROFILE% and create a .gemini folder.

2. Within the folder, create a file named settings.json.

3. Copy and paste the following into the settings.json file. Replace EMAIL with your Connect AI email and PAT with the PAT you obtained in the prerequisites:Copy and paste the following into the settings.json file. "Authorization" should be set to "Bearer OAUTH_JWT_TOKEN". This is the OAuth JWT token from the prerequisites.

   {
   "mcpServers": {
      "connect-cloud": {
         "httpUrl": "https://mcp.cloud.cdata.com/mcp/",
         "headers": {
           "Authorization": "Basic EMAIL:PAT"  //Replace with your Connect AI's Email and PAT
        },
         "trust": true
      }
    },
   "selectedAuthType": "oauth-personal"
   }   
   

4. In the terminal, type gemini. You may need to give Google authorization to access your account.

5. The Gemini screen appears as follows:

   

6. Enter a prompt in the prompt box to ask Gemini a question. Gemini uses the Connect AI tools such as getCatalogs, getSchemas, and getTables to retrieve data from connected data sources.

7. Gemini returns a response to your prompt.

   

Connect through the Google ADK

1. Download the folder adk-mcp-client and unzip the folder.

2. Rename .env example to .env.

3. In the .env file, replace your_email@example.com with your Connect AI email, and replace your_mcp_password_here with the PAT created in the prerequisites.In the .env file, remove MCP_USERNAME and MCP_PASSWORD. You are using the OAuth JWT token from the prerequisites to authenticate.

4. Run pip install -r requirements.txt in your terminal.

5. Run python -m google.adk.cli web --port 5000. This starts the ADK at port 5000 on your local machine.

6. In your web browser, open 127.0.0.1:5000. The application looks like the following:

   

7. Enter a prompt in the Type a Message box. Gemini uses the Connect AI tools such as getCatalogs, getSchemas, and getTables to retrieve data from connected data sources.