Skip to content
All posts

Tutorial · AI

Deploy AI Agents with n8n + OpenRouter on Hostendor

Run autonomous AI workflows 24/7 on Hostendor. Connect n8n to OpenRouter, build agent pipelines, and try it all on the 14-day trial.

AI agents are everywhere in 2026. But running them costs money — you need a server that stays on 24/7, and most platforms either charge for uptime or put your service to sleep.

Hostendor apps don't sleep. Ever. Here's how to run an AI agent workflow that checks your email, summarizes your day, and posts to Slack — and you can build it on the 14-day trial without adding a card.

What you'll build

An n8n workflow that:

  1. Wakes up every hour via a cron trigger
  2. Calls OpenRouter's API with a prompt
  3. Takes the AI's response and posts it to a webhook (Slack, Discord, Telegram — your choice)

Step 1: Deploy n8n from our template

Go to hostendor.com/templates and pick n8n. It typically deploys in 1-3 minutes. You'll get a URL like https://<id>.app.hostendor.com. Set your admin password and you're in.

n8n stores its workflows in a persistent volume, so your automations survive restarts and redeploys.

Step 2: Get an OpenRouter API key

OpenRouter gives you access to 200+ models through one API. Sign up at openrouter.ai, grab your key, and add $5 in credits (you'll use maybe $0.02/month for lightweight agents).

Step 3: Build the workflow

In n8n, create a new workflow with three nodes:

Cron trigger: Set to fire every hour. This is your agent's heartbeat.

HTTP Request node: POST to https://openrouter.ai/api/v1/chat/completions with headers Authorization: Bearer <your-key> and a JSON body containing your prompt. Use a lightweight model like google/gemma-4-26b-a4b-it:free — it's literally free on OpenRouter.

Webhook node: Post the AI's response to your Slack incoming webhook, Discord channel, or Telegram bot.

Step 4: What your agent can do

Once the plumbing works, the possibilities expand fast:

  • Daily briefing: Summarize your calendar, unread emails, and GitHub notifications into one Slack message every morning.
  • Content pipeline: Generate tweet drafts from your blog RSS feed, post them to a review channel.
  • Monitoring: Check your website's health endpoint and DM you if it's down.
  • Code review: Feed PR diffs to a coding model and post the feedback as a comment.

Why this works on Hostendor

Most platforms would charge you for the uptime, the compute, or the egress, or put your service to sleep. On Hostendor, n8n stays on 24/7 with no cold starts, and it runs comfortably on App Starter resources (1 vCPU, 1 GB). The trial's ₹300 of credit keeps one app running for the full 14 days, and OpenRouter's free models cost you nothing.

When the trial ends, App Starter is a flat ₹599/month. No metering, no surprise bills.

Start at hostendor.com/templates — pick n8n, deploy in one click.

Hostendor

Want to try this on your own project? The 14-day trial comes with ₹300 of credit, and you don't need a card to start.

More from the blog