weeny.cloud vs a plain VPS

This is the fairest comparison to make, because underneath, weeny is a VPS. One persistent Linux server, root, SSH, systemd, your own disk. If you're already comfortable running a box, you could rebuild most of weeny yourself in a weekend.

The question is whether you want to, and whether you want your agent to.

What you get on both

Root. SSH. systemd. A real filesystem that persists. Any language, any framework, any package, any binary. Long-running processes, cron, workers, databases. No platform ceiling.

If that's the whole list of what you need, rent a VPS — several are cheaper than weeny and you'll be fine.

What weeny adds

Everything below is the stuff you'd otherwise do by hand, and more importantly, the stuff your agent would otherwise do by hand — badly, at 2am, without telling you.

• A public HTTPS URL, instantly. weeny expose myapp 3000 and you have https://myapp-x7k2.onweeny.com. No nginx config, no certbot, no DNS record, no renewal cron that silently dies in ninety days.
• Process supervision that survives things. weeny start myapp -- npm start writes a systemd unit that restarts on crash and comes back after reboot. Your agent doesn't have to learn unit-file syntax to keep an app alive.
• Getting code onto the box. npx weeny-cloud push ./myapp — an agent-safe transfer that respects .gitignore and never makes the agent fight ssh keys or raw rsync flags.
• Off-machine backups and rebuild. The whole machine is snapshotted and backed up every few hours. If it breaks, one button restores it — apps, data, same URLs.
• Private apps with real identity. weeny expose --private plus weeny allow myapp [email protected] puts an auth gate in front of an app that has no auth of its own, and hands your app the visitor's email in X-Weeny-User. Doing this yourself means an OAuth proxy and an afternoon.
• Access for agents, not just browsers. npx weeny-cloud token <host> mints a bearer so another agent can call your private app as the person it was shared with. That one is genuinely hard to rebuild.
• Docs written for the agent. /llms.txt, a Claude Code skill, and worked recipes for Next.js, Python, Postgres, Docker and n8n. Your agent reads how the machine works instead of guessing.

The real argument

A VPS assumes a human operator who knows what they're doing. Every sharp edge — TLS, unit files, reverse proxies, firewall rules — is fine when a person handles it, and a reliable source of confidently wrong work when an agent does.

weeny is the same machine with those edges already handled, and with the remaining surface documented in a form an agent reads correctly. You keep the escape hatch: it's still Linux, you still have root, and if you want to ignore every weeny command and configure nginx yourself, nothing stops you.

When to just get a VPS

• You need a specific provider, region, or instance type weeny doesn't offer.
• You want to run it yourself, and enjoy it. That's a real answer.
• You need multiple servers, load balancing, or a private network between machines.
• You're price-sensitive at the very bottom end and don't need any of the above.

When weeny is worth it

You're handing the machine to a coding agent and you want it to ship working software rather than a half-configured web server — at a flat price, from $3/mo, with unlimited apps on the box.