Proxmox AI agent with n8n and generative AI integration

# Proxmox AI Agent with n8n and Generative AI Integration

This n8n workflow demonstrates how to create an AI agent that can interact with a Proxmox server, potentially through an HTTP API, and respond to user queries received via Telegram or Gmail. It leverages Langchain AI agents for intelligent decision-making and action execution.

## What it does

This workflow orchestrates an AI agent to perform actions based on incoming messages:

1.  **Receives Input**: It can be triggered by new messages from either:
    *   **Telegram**: Listens for incoming chat messages.
    *   **Gmail**: Monitors for new emails.
    *   **Webhook**: Can also be triggered by a generic HTTP POST request.
2.  **Initial Message Processing**: A `Code` node processes the incoming message, likely extracting the relevant text for the AI agent.
3.  **AI Agent Decision Making**: The core `AI Agent` node, powered by a `Google Gemini Chat Model`, analyzes the user's request. It's configured with:
    *   **Structured Output Parser**: To ensure the AI's output adheres to a predefined JSON structure.
    *   **Auto-fixing Output Parser**: To attempt to correct any malformed output from the AI, ensuring robustness.
    *   **HTTP Request Tool**: The AI agent has access to an `HTTP Request Tool`, allowing it to make API calls to external services (e.g., a Proxmox API).
4.  **Conditional Logic**: An `If` node checks if the AI agent's response contains a specific keyword or condition (e.g., "Proxmox").
5.  **Proxmox Interaction (Conditional)**:
    *   If the condition is met, the `HTTP Request` node is executed, presumably interacting with the Proxmox API based on the AI's decision.
    *   If the condition is not met, the workflow proceeds without Proxmox interaction.
6.  **Response Merging**: A `Merge` node combines the outputs from the conditional branches, ensuring a unified flow.
7.  **Final Output**: The workflow ends, ready to process the next incoming message.

## Prerequisites/Requirements

To use this workflow, you will need:

*   **n8n Instance**: A running n8n instance.
*   **Telegram Account & Bot Token**: For the Telegram Trigger. You'll need to create a Telegram Bot and obtain its API token.
*   **Gmail Account**: For the Gmail Trigger. You'll need to configure Gmail credentials in n8n.
*   **Google Gemini API Key**: For the `Google Gemini Chat Model`.
*   **Proxmox Server**: Access to a Proxmox server and its API documentation if you intend to use the `HTTP Request` node for Proxmox interaction.
*   **API Endpoint for Proxmox (Optional)**: If the `HTTP Request` node is indeed used for Proxmox, you'll need the relevant API endpoint and authentication details.

## Setup/Usage

1.  **Import the Workflow**:
    *   Download the provided JSON file.
    *   In your n8n instance, go to "Workflows" and click "New".
    *   Click the "Import from JSON" button and paste the workflow JSON or upload the file.
2.  **Configure Credentials**:
    *   **Telegram Trigger**: Configure your Telegram Bot API credentials.
    *   **Gmail Trigger**: Configure your Google OAuth2 credentials for Gmail.
    *   **Google Gemini Chat Model**: Configure your Google Gemini API key.
3.  **Configure AI Agent**:
    *   Open the `AI Agent` node and ensure the `Google Gemini Chat Model` is correctly linked and configured with your API key.
    *   Review the `Structured Output Parser` and `Auto-fixing Output Parser` settings.
    *   Examine the `HTTP Request Tool` within the AI Agent to understand how it's intended to interact with external APIs. You might need to define its capabilities (e.g., "tool to manage Proxmox VMs").
4.  **Configure HTTP Request Node (Proxmox Interaction)**:
    *   Open the `HTTP Request` node (node ID 19).
    *   Set the `URL`, `Method`, `Headers`, and `Body` according to your Proxmox API documentation. This node will be executed if the AI agent determines a Proxmox action is needed.
5.  **Configure If Node**:
    *   Review the conditions in the `If` node (node ID 20) to ensure it correctly identifies when Proxmox-related actions should be taken based on the AI agent's output.
6.  **Activate the Workflow**: Once all configurations are complete, activate the workflow.

Now, when a message is received via Telegram, Gmail, or the Webhook, the AI agent will process it, decide on an appropriate action (potentially interacting with Proxmox), and follow the defined logic.