• EXECUTION
  • DATA
  • SECURITY
  • HUMAN

How to execute proprietary api calls with your AI

Connect your private endpoints to your AI client through a secure bridge. Your agent triggers internal logic without touching your raw credentials.

Ask AI about this page

Short answer

How can I let my AI call my private APIs?

You use this MCP to bridge the gap between your AI and Google Cloud. The agent triggers specific functions via a secure proxy, keeping your API keys and secrets isolated from the model. This setup gives you a controlled way to run internal logic through chat.

The outcomes.

Where your requests land.

The AI maps your natural language requests to specific cloud triggers.

EXECUTION

Cloud function triggers

The AI identifies which function to run based on your prompt. It sends the necessary payload to Google Cloud to execute the code.

DATA

Function responses

The output from your code flows back to the AI. You get the results of the API call directly in your chat interface.

SECURITY

Isolated credentials

All authentication happens in the background. Your AI never sees the keys used to authorize the call.

HUMAN

Manual oversight

You decide which prompts are sent to the cloud. The AI cannot modify your function code or change your cloud configuration.

The workflow

What your AI does when a request arrives.

The AI acts as the interface between your conversation and your cloud infrastructure.

  1. Parse the intent

    Your agent analyzes your message to determine which specific cloud function needs to run and what arguments it requires.

    invoke_function

  2. Securely bridge the call

    The MCP handles the heavy lifting of passing the request to Google Cloud using the pre-configured connection.

    invoke_function

  3. Process the response

    Once the function finishes, the AI receives the raw data and translates it into a readable answer for you.

    invoke_function

  4. Report results

    The AI presents the final outcome of the API call, such as a status update or retrieved data, in the chat.

    invoke_function

Try it

Copy these to start.

Use these prompts to trigger your functions.

Starting points

Replace the function names and parameters with your actual Google Cloud Function identifiers.

Google Cloud Functions Connector

You're all set. Choose your MCP client and follow the setup instructions.

Connector linkhttps://edge.vinkius.com/vk_preview_ot6QbSG45dUkURNWnf6YGaSHZvntFNxbPzDeJA49/mcp

Claude Desktop

Follow the steps below to connect in seconds.

  1. 1In Claude Desktop, open Settings → Connectors.
  2. 2Click “Add custom connector” and paste the connector link above as the remote MCP server URL.
  3. 3Click Add and start a new chat — Google Cloud Functions capabilities are ready to use.
Configuration · claude_desktop_config.jsonCopy
{
  "mcpServers": {
    "google-cloud-functions-mcp": {
      "url": "https://edge.vinkius.com/vk_preview_ot6QbSG45dUkURNWnf6YGaSHZvntFNxbPzDeJA49/mcp"
    }
  }
}
Copy into chat04

  • Run the 'get_user_status' function for user ID 5502.

  • Trigger the 'process_invoice' function with the data from this email.

  • Call the 'check_inventory' function for the item SKU-992.

  • Execute 'update_record' to set the status to 'completed' for entry 44.

  • Claude
  • ChatGPT
  • Cursor
  • VS Code
  • Windsurf
  • Claude Code
  • JetBrains
  • Cline

Start here

Connect Google Cloud Functions once, then ask.

Link your Google Cloud account in one step. Your credentials stay encrypted and managed by Vinkius, and you can find the full setup guide on the Connector page.

Connect Google Cloud Functions to your AI

FAQ

How this task behaves.

  • 01

    Does the AI see my Google Cloud service account keys?

    No. The keys are stored securely in the Vinkius environment and are never passed to the AI model.

  • 02

    Can the AI change my function code?

    No. The AI can only trigger existing functions using the invoke_function tool. It has no permission to edit your cloud resources.

  • 03

    Can I use this to call any API?

    You can call any API that is wrapped in a Google Cloud Function. The AI interacts with the function, not the raw endpoint.

  • 04

    What happens if the function fails?

    The AI receives the error message returned by the function and will report that failure to you in the chat.

  • 05

    Can the AI delete my functions?

    No. This MCP only provides the ability to invoke functions. It does not have tools for managing or deleting cloud infrastructure.

  • More questions about Google Cloud Functions? See everything the Google Cloud Functions Connector can do

Connect Google Cloud Functions to Claude, Cursor, ChatGPT & more