A lab of apps for $20.
Side projects should cost nerve, not rent. Run your whole portfolio — experiments included — on one dedicated compute where the 11th app costs the same as the 1st: nothing.
The indie math
- 1 project
- $5/mo
- 10 projects
- $20 flat
- Killed experiments
- $0 sunk
- Hook invocations
- Unbilled
Pro is one 2 vCPU compute, not a quota. Frontends are near-free static files, instances are small Go processes — backends fill the box first.
How a weekend ships
Local binary in, live URL out
01
Develop locally
pbc local initThe exact same PocketBase binary locally and in production. No dev/prod drift to debug on Sunday night.
02
Add logic in hooks
pb_hooks/*.pb.jsCustom routes, record events and cron next to your data — deployed with the instance, billed never.
03
Deploy the bundle
pbc deployFrontend + backend + DB from one CLI. Redeploys remember the target, CI ships via GitHub Action.
Why indies stay
One binary to understand fully
SQLite file next to the process. One log to read, one folder to back up, API rules as one-line filters.
Hooks replace a backend (until they don’t)
Most MVPs never need more. When they do, Pro runs Node, Next, Deno or Bun beside the instance.
Frontend included, not invoiced
5 frontends on Free/Starter, unlimited on Pro. No second dashboard for the landing page.
Kill apps without grief
Failed experiments delete in a click. Survivors already sit on infrastructure that scales vertically.
Honest fit check
Good fit
- Read-heavy CRUD: SaaS dashboards, tools, content apps, mobile backends
- Solo devs who want local == production and one CLI for everything
- Portfolios where flat pricing beats per-project multiplication
Not the fit
- High-concurrency write fan-out or Postgres extensions (PostGIS, pgvector) — consider Supabase
- Global-edge or scale-to-zero needs — consider Cloudflare or Vercel
Put your next project on PocketBase Cloud.
No lock-in: your data is a stock PocketBase directory you can download as a zip and restore anywhere, any time. How that works.