---
title: "Hostinger vs Vercel: what a Next.js app actually costs to host"
description: "Vercel bills you for what your Next.js app used; Hostinger bills you for the plan you picked. Here is the same app costed both ways from published rates — including where Vercel is worth every cent, and where a flat plan is the saner buy."
canonical: https://wacrm.tech/blog/hostinger-vs-vercel
datePublished: 2026-08-25
dateModified: 2026-08-25
author: "Arnas Donauskas"
license: "content © wacrm; product is MIT-licensed"
---
# Hostinger vs Vercel: what a Next.js app actually costs to host

Vercel and Hostinger do not really compete on price. They compete on *how* you
pay. Vercel meters what your app used and sends an invoice; Hostinger sells you
a plan and the invoice is the plan. Most of the frustration people report with
hosting bills is that difference, not the numbers.

So this comparison does two things. It costs the same Next.js app both ways
using each vendor's published rates, and it says plainly where the metered model
is worth paying for — because for some apps it clearly is.

> **Disclosure:** we publish this, we host [wacrm](https://wacrm.tech/) on Hostinger, and the
> Hostinger links here are referral links. Every figure below is calculated from
> the vendors' own pricing pages, checked in August 2026, and the arithmetic is
> shown so you can check it. Prices change; verify before you buy.

## The short version

- **Choose Vercel** if traffic is spiky, your visitors are spread across
  continents, or your team lives on preview deployments and instant rollbacks.
  You are buying a global edge network and a release workflow, and it is a good
  one.
- **Choose Hostinger** if you want the hosting line in your budget to be a
  number you chose. A managed Node.js plan is $3.99/month on term ($16.99 at
  renewal) and runs five apps with unlimited bandwidth, real cron, and a
  persistent Node process.
- **The floor matters more than the ceiling.** Vercel's free Hobby plan is not
  available for commercial use, so a business app starts at $20/month however
  quiet it is.

## What it costs, both ways

Pick the traffic level that looks like yours. The Vercel column is built from
its published Pro rates with the included allowances and the $20 monthly credit
applied — the two things most "Vercel is expensive" posts leave out.


The pattern is worth stating out loud: at small and medium size, **almost all of
a Vercel bill is seats, not usage**. Three developers who can each press deploy
is $60/month before a single visitor arrives. Usage only takes over once you
pass the allowances, and then it moves quickly.

## What Vercel meters

| Resource | Included on Pro | Rate after that |
| --- | --- | --- |
| Deploying seats | 1 | $20/month each |
| Fast data transfer | 1 TB | from $0.15/GB |
| Edge requests | 10M | from $2.00 per million |
| Function invocations | — | $0.60 per million |
| Active CPU | — | from $0.128/hour |
| Provisioned memory | — | from $0.0106/GB-hour |
| Build compute | — | $0.0035 per CPU-minute |
| Image optimization | — | $0.05 per 1,000 transformations |
| Observability Plus | — | $1.20 per million events |

Two footnotes that matter. The "from" prices are the cheapest US region; serving
from Frankfurt, Mumbai or São Paulo costs more per unit for the same traffic.
And the $20 monthly credit offsets usage only — never the platform fee or seats,
and it does not roll over.

None of this is hidden. It is published in detail, which is more than many hosts
manage. The problem is not disclosure, it is that eight metered dimensions make
next month's bill genuinely hard to predict — and Vercel's own default spend
notification sits at $200 per cycle, which tells you something about the range
they expect people to land in.

## The Hobby plan is not a business option

This is the part most comparisons miss, and it changes the maths more than any
unit rate. Vercel's fair use guidelines restrict Hobby teams to non-commercial
personal use, and define commercial usage broadly: taking payment, advertising a
product or service, being paid to build or host the site, running ads — even
asking for donations. All of it requires Pro or Enterprise.

So "Vercel is free" is true for a portfolio and false for a business. The
realistic comparison is $20/month against $3.99–7.99/month, before usage.

## What the flat plan buys


Beyond the specs in those cards, every plan includes a first-year domain,
unlimited SSL certificates, backups, and 24/7 human support. A Git integration
builds and deploys from your GitHub repository on push, so the everyday loop is
the same one you are used to: commit, push, watch it deploy.

What you are **not** buying, and should not pretend you are:

- **A global edge network.** Your app serves from one data centre. Pick the
  region nearest your users and accept that visitors on the other side of the
  world will feel it.
- **Autoscaling.** A traffic spike does not create an overage line; it creates
  slower responses until you move up a tier. That is a real trade, not a free
  win.
- **Preview deployments per pull request.** You get one deployed app per app
  slot, so staging is a second slot you set up yourself.

## Where Vercel is clearly worth it

An honest comparison has to include the cases where you should pay Vercel and
stop reading:

- **Spiky or unpredictable traffic.** Launches, campaigns, anything that can go
  from a hundred to a hundred thousand visitors in an hour. Autoscaling you do
  not have to think about is worth real money.
- **A globally distributed audience.** Serving from the edge in dozens of
  locations is not something a single shared container can imitate.
- **Teams that ship on previews.** A URL per pull request, comments on the
  deployment, one-click rollback — this genuinely changes how a team reviews
  work, and rebuilding it elsewhere costs more than the subscription.
- **Vercel-specific features.** Edge Config, skew protection, built-in
  analytics and observability, the AI SDK integrations. If you use them, you are
  not comparing hosts; you are comparing a host with a platform.

## Where the flat plan wins

- **A bill you can forecast.** The number is the same every month. For a
  bootstrapped product or a client site, that is often worth more than a lower
  average.
- **A persistent Node process.** A long-running send loop, a queue worker, or a
  slow import runs in the process instead of being chopped up to fit inside a
  function's execution limit.
- **Ordinary cron, at any interval.** On Vercel, cron jobs invoke functions, and
  the schedule depends on the plan: once per day on Hobby (with up to an hour of
  slack), once per minute on Pro. On a Node.js plan you get normal cron entries
  at whatever interval you want.
- **Several apps on one plan.** Business runs five web apps and 50 websites. On
  a metered platform every extra project is another set of meters.
- **Unlimited bandwidth.** Traffic that would cost $161 in transfer overage at
  2 TB a month costs nothing extra here. This is the single biggest structural
  difference between the two models.

## Running Next.js on a Node server

Worth being precise, because there is a lot of folklore about this. A managed
Node.js plan runs a real Node process, so you build the app and serve it with
`next start` — the framework's own production server. Server components, route
handlers, ISR, `next/image` optimization, middleware and streaming all work as
documented, because they are framework features, not platform features.

What you lose is the Vercel-specific layer: their analytics and observability,
Edge Config, per-PR preview URLs, and edge-runtime middleware running in dozens
of locations rather than in your one Node process. What you gain is a machine
that stays up between requests, with a filesystem and a crontab.

## Moving an app off Vercel

A realistic afternoon's work for a straightforward app:

1. **Check the build locally.** `npm run build && npm start` on the Node version
   your plan offers. Note the peak memory — this is what decides your plan tier.
2. **Create the app.** Buy a plan, create a Node.js application in hPanel, set
   the start command to `npm start`.
3. **Connect the repository.** Point the Git integration at your GitHub repo and
   branch, then copy your environment variables across. This is the step people
   forget: anything read at build time has to exist before the first build.
4. **Move the domain.** Point DNS at the new host and let managed SSL issue a
   certificate. Keep the old deployment live until the new one answers.
5. **Recreate scheduled work.** Vercel cron entries in `vercel.json` become cron
   jobs in hPanel hitting the same routes.
6. **Consider keeping Vercel for previews.** A Hobby team for personal preview
   deploys plus production on a flat plan is a legitimate combination, provided
   the previews themselves are not commercial use.

## If the app is wacrm

That is what this site runs, and the [Hostinger deployment
guide](https://wacrm.tech/docs/deployment-hostinger) is the step-by-step version: connect your
fork, set the Supabase and Meta environment variables, and the CRM is live. The
Business plan comfortably runs the app plus a marketing site, and automations use
real cron rather than a once-a-day approximation.

For the wider argument about self-hosting a CRM rather than renting one, the
[free and open-source WhatsApp CRM guide](https://wacrm.tech/blog/free-open-source-whatsapp-crm)
covers the total cost of ownership.

## The honest bottom line

If your app is quiet, Vercel Pro at $20/month is not expensive, and the platform
is genuinely excellent — nobody should switch hosts over $16. Switch when one of
these becomes true: usage has started to appear on the invoice, you are paying
for seats rather than traffic, you want work that does not fit a function, or you
simply want to stop watching a meter.

If that is where you are, the
[Business plan](https://www.hostinger.com/web-apps-hosting?REFERRALCODE=WACRMHOST)
is the one to start with, and there is a 30-day money-back guarantee if the fit
turns out to be wrong.