Turn chat, email, and web requests into safe Proxmox actions. Teams can ask in plain language to start or stop VMs, check cluster status, or change settings. Great for IT teams that want faster operations and less time in the UI.

Messages from Telegram, Gmail, a chat interface, or a webhook feed into an AI agent. The model reads your request along with Proxmox API docs and wiki to plan the right call. A structured parser forces a clean JSON output with method, URL, and body fields. Switch and If nodes route that request to the correct Proxmox HTTP endpoint with a secure header token. Results are merged, cleaned, and explained in simple words so anyone can understand the outcome.

Setup needs a Proxmox API token with proper rights and a Google Gemini API key. Update the base API URL to your host and pick your model. Expect faster ticket handling, fewer console logins, and clear audit trails. Ideal for chat based help desk, weekend duty handoffs, or safe self service for trusted users.

• Multi channel intake with Telegram Trigger, Gmail Trigger, chat trigger, and Webhook.
• AI agent uses Google Gemini plus Proxmox API docs and wiki to plan the right API call.
• Structured Output Parser and auto fixing parser enforce valid method, URL, and body fields.
• Switch and If nodes route to GET, POST, PUT, DELETE, or PATCH HTTP Request nodes.
• Secure header auth with PVE API token and optional support for self signed certificates.
• Structure Response and Format Response and Hide Sensitive Data clean and mask outputs.
• Merge nodes join multiple branches so replies contain the right combined result.
• Code node wraps the final data into a clear and readable message.

• Reduce manual VM tasks from 30 minutes to under 2 minutes by using chat or email commands.
• Automate up to 80 percent of routine server requests with clear routing and checks.
• Cut command errors by 90 percent through a strict structured output parser.
• Connect chat, email, and webhooks in one place for server actions.
• Support five times more requests per engineer with the same team size.

1. Import the template into n8n: Create a new workflow in n8n > Click the three dots menu > Select 'Import from File' > Choose the downloaded JSON file.
2. You'll need accounts with Proxmox VE, Google Gemini, Telegram and Gmail. See the Tools Required section above for links to create accounts with these services.
3. In your Proxmox UI create an API token in the Data Center. Build the header value as PVEAPIToken=user@realm!tokenid=token. In the n8n credentials manager create an HTTP Header Auth credential named Proxmox with header name Authorization and paste the token value.
4. Open each HTTP Request node and set the base URL to your Proxmox host using the path /api2/json. Keep allowUnauthorizedCerts only if your host uses a self signed certificate.
5. Double click the Gmail Trigger node. In the credential dropdown choose Create new credential and follow the on screen steps to sign in to your Google account and approve Gmail API access.
6. In Telegram create a bot with BotFather and copy the bot token. Open the Telegram Trigger node, create a new Telegram credential, paste the token, and save.
7. Open both Google Gemini Chat Model nodes. Choose Create new credential for Google Gemini, paste your API key from your Google AI account, and select the model you want to use.
8. Open the Webhook node and copy the test URL for inbound requests. If you want a secret header, set it in the node and send the same header from your client when posting to the webhook.
9. Review the Structured Output Parser node. Confirm the sample schema matches your Proxmox paths and update any node names or VM ids used in examples.
10. Enable the workflow. Send a Telegram message like start VM 105 or email the Gmail inbox with a simple request. Watch the execution and confirm the Proxmox action and the human friendly reply.
11. If you see 401 errors check the token format and permissions. For TLS issues enable allowUnauthorizedCerts only if needed. If the parser fails, simplify your prompt and try again.