DEV Community

David Bos
David Bos

Posted on

Ship on a free link first, buy the domain when it has visitors

#sideprojects #webdev #startup #productivity

Disclosure: I'm behind DropTheHassle.

In September someone posted this on Hacker News: "I'm addicted to buying domains for my vibe-coded apps that get 0 visitors". The post had no text, just the title. It didn't need any. The comments read like a support group.

"I feel so bad when I look at my domains after few months lol." (arpanghoshal, the poster)

"It's not you, it's all of us" (wakaru44)

"I have porkbun open in another tab." (jtap)

"I'm nearly up to 100" (alexnewman)

I build a tool that sells domains, so I have an obvious interest here. But I think the thread gets to something real about how side projects are built now, and the advice in it was better than "just stop".

Why we buy the domain first

When an AI can build a working app in an afternoon, the domain becomes the most "real" thing you can do next. It costs money, it has your idea's name on it, and it feels like commitment.

One commenter named it:

"Buying a domain is pretend work. It feels like progress, but doesn't make a project/product better…" (brudgers)

Another, more fun, version:

"buying domains that just seem fun or cool" [...] "What have I done with it? Nothing." (davidsojevic)

The problem isn't the one-time cost. It's the renewals. A year later the project is dead, the domain is still charging you, and every renewal email is a small reminder.

Camp 1: use the free link

The most common advice was to not buy anything until you have to:

"Get a single domain and use subdomains till one of your projects pop off. Or use the cloudflare worker/vercel domain that is by-default provided" (datron)

"someshit.vercel.app bridges the gap." (nsnsnsnsjsj)

"If you get one good domain, subdomains work really well" (toast0)

That's sound. A free platform link (*.vercel.app, *.pages.dev, *.netlify.app, *.github.io) is plenty to find out whether anyone cares.

Camp 2: keep it clean from day one

One comment pushed back, and it's worth reading in full:

"if one of the side-projects takes off, it's a real pain to disentangle… migrating a project to its own domain all to save less than $100… keep your projects clean from the get go" (takinola)

This is the part people forget. If project #7 lives at project7.yourname.com, then its links, its search history, its users' bookmarks, maybe its cookies and its email all hang off your personal domain. Moving it later is a migration, not a rename.

Both camps are right, so combine them

The rule I use, and the one I'd suggest:

Ship on a free link first. Buy the domain when it has visitors. But give every project its own free link, not a subdomain of your personal domain.

That keeps the cost at zero while you find out whether it matters, and it avoids the takinola problem: each project is already its own site. When one earns a name, you point a domain at that site. You don't untangle it from everything else.

What counts as "has visitors" is up to you. Mine is roughly: someone I didn't send it to used it, or I'm about to put the link somewhere permanent (a bio, a client proposal, a launch post). Before that, the free link is enough.

Make the free link a real deploy

The failure mode isn't only buying domains too early. It's also never shipping at all, because "I'll deploy it properly once I have the domain."

Whatever host you use, get it on a public URL the day you build it. Open it on your phone and send it to one person. That's the test that matters.

If you build with Claude Code, Cursor, Codex or Windsurf, here's the version I built, which needs no account:

npx -y dropthehassle deploy
Enter fullscreen mode Exit fullscreen mode

It uploads the built folder and prints a live name.dropthehassle.app link with HTTPS, plus a claim link. Your agent can also do it through an MCP server (https://dropthehassle.com/mcp) with no token. The honest details:

  • the claim link works for 7 days; an unclaimed site is removed after that;
  • until you claim it, the site is served with noindex and shows a small, dismissible "Claim this site" bar, so it's a preview, not a launch;
  • static sites only (HTML or a build output), no server code.

When it does earn a domain

Two things to check before you buy, wherever you buy:

  1. The renewal price, not the first-year price. Plenty of registrars discount the first year. The renewal is what you'll actually live with, and it's the number behind the guilt in that thread.
  2. Whether you'll need to touch DNS. If you buy the domain where the site already lives, you often don't. If you buy it elsewhere, you're adding records at the registrar.

For what it's worth, on DropTheHassle a .com is €19 a year including VAT in the EU, or $19 elsewhere, and it's the same every year. If you buy it for a site that's already there, it goes live on that site after payment, with no DNS records to edit. A domain you already own needs one A record. And if you let your AI handle it: it can search for names and publish the site, but it can never pay. The checkout is always your click.

The short version

  • Build it, deploy it the same day, on a free link.
  • Give each side project its own link, not a subdomain you'll have to untangle.
  • Buy the domain when the project has visitors, and check the renewal price when you do.

And if you already have a drawer of domains with nothing on them: as one commenter put it: "ARE U ME" (wompapumpum). You're not the only one.

If you want the step-by-step for a specific tool: guides for Claude Code, Cursor, Lovable and more.

Disclosure: I'm behind DropTheHassle.

Top comments (0)