xhostd
Sign in Start building
← Solutions

Run a bot or data tool on xhostd

Run a chat bot or data tool on xhostd: a background worker that keeps running, with Postgres, secrets, logs, and a predictable monthly price.

A background worker is an app that keeps running even when nobody visits a web page. On xhostd, your AI coding agent deploys your bot or data tool as a background worker, with its data in Postgres and its tokens stored as secrets.

Who this is for

You want a tool that keeps working while nobody looks at it. For example:

  • a Discord or Slack bot, such as one that posts daily stats to your server
  • a data collector or monitor, such as a script that checks a supplier's price list every 10 minutes, or a website every hour, and records the changes
  • an AI processing worker that summarizes or sorts new items as they arrive

How it works

  • A worker runs a loop instead of serving a page. It listens to a chat service, or it wakes up every few minutes to do its job.
  • Your agent builds and deploys it. The worker runs in a channel, one running copy of your app.
  • A tool with a dashboard runs as two apps. One is the worker, and the other is the web app that people open in a browser.
  • Lasting data goes in the database. Anything that the worker must keep goes in its Postgres database or its file storage.
  • Test changes first. Ask your agent to deploy each change to a preview channel first. Read Preview and staging environments on xhostd.

Give your agent a clear brief. For example:

Build a Python worker that checks a data source every 10 minutes,
stores new rows in Postgres, and logs one line per check.
Deploy it to xhostd as a background worker.

What xhostd provides

  • Always running. The worker never sleeps, on every plan.
  • A database. Each channel that runs code has its own Postgres database.
  • Secrets. xhostd encrypts each token and key, and it gives them to the worker when it starts. Your agent can set a secret, but it can't read one back unless you allow that in the console.
  • Logs. Your agent reads the worker's log. The console shows the recent part, along with health findings and your usage against your plan.
  • Predictable pricing. Nothing is charged per request, so a busy day doesn't raise the price. Start on the free plan, watch your usage for a few days, and upgrade when you reach a limit. Read the pricing page.

Next step

When you're ready, give your agent the worker recipe. If you haven't connected an agent yet, start by connecting your agent.

Technical notes

  • Secrets in the console. When you add a token or key in the console, select Store as a secret. A plain setting is the default kind, and your agent can read it back, so never store a token as one.