AI Proxmox Management Agent with n8n and Google Gemini

Unlock Intuitive Proxmox VE Management with this AI Agent

Tired of manually crafting Proxmox API calls or navigating complex UIs for routine tasks? This AI Agent acts as your intelligent Proxmox assistant, translating your natural language requests into precise API commands. Powered by Google Gemini and n8n, it allows you to manage virtual machines, nodes, and other Proxmox resources conversationally.

Whether you're a solopreneur managing a few VMs or a CTO looking to empower your team, this agent streamlines Proxmox operations, saving you time and reducing the learning curve.

Key Features & Benefits

• Natural Language Control: Simply tell the agent what you want to do (e.g., "Create a new Ubuntu VM with 2 cores," "List all VMs on node psb1," "Delete VM 105").
• AI-Powered Command Generation: Google Gemini interprets your intent and, guided by integrated Proxmox documentation tools, constructs the correct API requests (GET, POST, PUT, DELETE).
• Automated Proxmox Interaction: The agent directly executes commands on your Proxmox server(s).
• Structured Output: Ensures API calls are correctly formatted in JSON and handles responses.
• Response Interpretation: A secondary AI agent helps summarize Proxmox GET request outputs into human-readable text.
• Flexible Triggers: Initiate commands via n8n's built-in chat, Telegram, Gmail, Webhooks, or integrate with your existing systems.
• Knowledge-Augmented: The AI Agent uses Langchain tools to consult Proxmox API documentation and your live cluster status for more accurate command generation.
• Error Handling & Auto-Fixing: Includes an auto-fixing output parser to improve the reliability of AI-generated JSON.

This AI Agent empowers you to manage your Proxmox VE environment more efficiently and intuitively, making complex infrastructure tasks accessible through simple conversation.

• An n8n instance (Cloud or self-hosted).
• Proxmox VE instance(s) accessible from your n8n instance.
• A Proxmox API Token. See the workflow's sticky note for guidance on creating one (Data Center > API Tokens in Proxmox).
• Google Cloud Platform account with the Vertex AI API enabled.
• Credentials for Google Gemini (API Key or Service Account, configured in n8n).
• (Optional) Credentials for your chosen trigger (e.g., Telegram Bot API Key, Gmail OAuth2).

1. Download the n8n workflow JSON file.
2. Import the workflow into your n8n instance.
3. Create and configure Proxmox API Token credentials in n8n: Navigate to 'Credentials' in n8n, click 'Add credential', search for 'Header Auth'. Name it (e.g., 'Proxmox API Auth'). For 'Name', use Authorization. For 'Value', use PVEAPIToken=YOUR_USER@YOUR_REALM!YOUR_TOKENID=YOUR_TOKEN_SECRET. Assign this credential to the 'Proxmox' tool node and all 'HTTP Request' nodes interacting with Proxmox.
4. Create and configure Google Gemini credentials in n8n: Navigate to 'Credentials', add 'Google Gemini (PaLM) API'. Enter your API Key. Assign this to all 'Google Gemini Chat Model' nodes.
5. Review and update Proxmox URLs: In the 'Proxmox' tool node (ID 80b1ef4d-b4c7-40b4-969f-f53d0068cac7) and all 'HTTP Request' nodes, ensure the base URL (e.g., https://10.11.12.101:8006/api2/json) matches your Proxmox server's address.
6. Customize the main 'AI Agent' (ID 1492e53e-66b5-485b-b7e5-a42b76ebccb6): Review its prompt. Update default Proxmox node names (e.g., 'psb1') if yours are different. The agent is pre-configured with examples like creating, deleting, starting, stopping VMs.
7. (Optional) Configure your desired trigger node (e.g., 'Telegram Trigger', 'Gmail Trigger', 'Webhook'). The 'When chat message received' trigger is enabled by default for easy testing directly in the n8n UI.
8. Activate the workflow.
9. Test by sending a command through your chosen trigger, e.g., "List all VMs on psb1" or "Create a Debian VM with 2 cores and 2GB RAM".