Vercel alternatives.
Four honest options.
Most teams start looking because the bill grew in a shape they didn't expect: metered usage on one line, and a separate database subscription on another. These are the realistic paths out, with trade-offs stated plainly. One of them is ours.
Four honest choices.
We build one of these, so read with that in mind. Trade-offs are listed for all four, ours included, and two of them are things we genuinely can't do.
Sproobo
Flat bill, your server, databases included.
Strengths
- No per-database price: services are containers on hardware you own
- Flat plan per organization, so the bill doesn't move with traffic
- PR previews with their own database, on your own box
- Data on plain host paths you can back up and take with you
Trade-offs
- No global edge network: one origin, or your own CDN
- No burst autoscaling: your server size is the capacity
- Edge and serverless-only APIs don't carry over
Cloudflare Workers
Stay serverless, on the cheapest bandwidth in the business.
Strengths
- Generous free tier and famously inexpensive egress
- Genuine edge execution with a large global network
- D1, KV and R2 cover many apps without a separate DB vendor
Trade-offs
- The Workers runtime is not plain Node, so some libraries don't run
- Still usage-metered, so the bill still moves with traffic
- Porting a Next.js app can mean real rework
Coolify
Open-source, self-hosted PaaS on your own hardware.
Strengths
- Free and open source, with a large template catalog
- Containers on servers you own, so no usage meter at all
- Nothing managed by a vendor, if that is your requirement
Trade-offs
- You operate the platform: upgrades, its backups, its restores
- Durable data defaults to named volumes rather than host paths
- Builds run on the same box that serves production
Roll it yourself
A VPS, Docker and a proxy, wired by hand.
Strengths
- Total control and zero platform fees
- No usage meter anywhere in the stack
- You learn every layer you depend on
Trade-offs
- Deploys, TLS, rollbacks, backups and alerting are all yours to build
- Every server drifts into a snowflake without automation
- Bus factor of one
Count the subscriptions, not the platform.
Before comparing platforms, add up what sits beside the platform. A typical stack carries a Neon, Supabase, PlanetScale or Upstash subscription next to it, often one per environment, and that half of the bill is easy to overlook when you only compare the headline price.
Ask what a second database costs
On usage-billed stacks it is another subscription or another metered project. On a server you own it is another container, which is a click rather than a purchase.
Ask what staging costs
A full staging copy multiplies every metered line at once. If environments are the thing you ration to control spend, that is the meter talking, not your architecture.
Ask what leaving costs
Data spread across a platform and two vendors is an export project. Data on host paths on your own box is a directory you already control.
Sproobo vs Vercel, in detail.
Green on both sides where it's deserved. Vercel does things we structurally can't, and the table says so.
| Sproobo | Vercel | |
|---|---|---|
| Where your app runs | Your server, in Docker | Vercel's managed cloud |
| Cost shape | Flat: plan + your server | Per seat + metered usage |
| Databases | Containers, no per-DB fee | Separate paid products |
| A second database costs | Nothing new | Another subscription |
| Bandwidth | Whatever your host includes | Metered past allowance |
| Burst autoscaling | No, your server is the capacity | Yes, serverless |
| Global edge network | No, one origin or your own CDN | Yes, built in |
| PR preview deployments | Built-in, on your server | Built-in |
| Where your data lives | Your box, on host paths | Spread across vendors |
| If you leave | Stack keeps running as Docker + Caddy | Re-platform functions, move data out |
Vercel capabilities and pricing structure described from public documentation and change over time. Spotted something out of date? Tell us and we'll fix it.