Supabase Alternative, The PocketBase Empire Stack

What you'll build

A clear understanding of when PocketBase wins and when Supabase wins, the cost math at three user-volume tiers, and a decision matrix you can apply to your project. Roughly 12 minutes to read, no install required.

Caption: PocketBase admin UI on the left, Supabase Studio on the right, both managing user collections.

Prerequisites

• A clear sense of your actual user volume (not your fantasy user volume)
• Comfort with the trade-off "single binary on my VM" versus "managed Postgres with row-level security"

This is a decision tutorial, not an install guide. The install for PocketBase is in another tutorial in this set.

Step 1, the licence and pricing math

The Supabase Pro tier kicks in at the moment your DB exceeds 500MB or you cross 50k monthly active users. For an empire-scale solo SaaS, that takes years to hit; you might never. PocketBase on Oracle ARM is genuinely free.

Step 2, the feature parity check

For most Indian SaaS use cases, the feature surface area is similar. The difference is "Postgres" versus "SQLite" and "managed" versus "self-hosted".

Step 3, when Supabase wins

Supabase is the right pick when:

• Your team has Postgres expertise but no Go expertise
• You need horizontal scaling past 100k DAU
• You want managed backups, point-in-time recovery, replication
• You are a venture-backed startup and "self-hosting infra" is a distraction
• Compliance requires US/EU data residency with managed audit logs

For these cases, the $25-599/mo is well-spent.

In any of these, switching to PocketBase to save money is a false economy.

Step 4, when PocketBase wins

PocketBase is the right pick when:

• You are a solo operator or small studio building empire-scale SaaS
• Your DAU ceiling is realistically under 10k
• You want true zero-rupee fixed cost (only Anthropic API spend)
• You can write or have written ~50 lines of Go for custom routes
• You value single-binary simplicity over Postgres's feature breadth

For these cases, PocketBase plus Oracle ARM is decisively better.

For my empire (six SaaS apps, none with more than 10k MAU), PocketBase wins on every project I evaluate.

Step 5, the migration math

If you started on Supabase and want to switch to PocketBase, the migration costs:

• Schema port: ~4 hours per moderately complex schema
• Auth migration: ~6 hours including testing
• Custom routes (if any): ~1 hour per Go route
• Frontend client swap: ~2 hours for stock React, more for SSR
• Total for a typical 10-table SaaS: 20-30 hours of focused work

Versus the savings of $25-50/mo, the payback is 2-4 months for a solo operator. For a team paying engineer time, the math is less obvious.

First run

A real decision I made for a recent client project:

Project: Indian SaaS for legal document automation
Expected DAU: 200-2000 (year 1)
Compliance: India-only data
Budget: Bootstrap-friendly

Decision: PocketBase plus Coolify plus Oracle ARM
Reason: Free tier covers expected scale; India-region Oracle data centre; 
  PocketBase admin gives the client a self-service workflow without an
  engineer involved.

Outcome: Live for 4 months, Rs 0/mo fixed cost. Anthropic spend tracks usage.

The decision was decisive on the cost angle plus the data residency angle.

What broke for me

Two specifics from past migrations. First, Supabase's row-level security is materially more flexible than PocketBase's rules engine. Migrating a complex RLS policy to PocketBase rules took me a full afternoon for one collection because the SQL-like flexibility of RLS does not have a one-to-one mapping. The fix was simplifying the policy at the application layer; the trade-off was real.

Second, Supabase's storage signed-URL flow is more polished than PocketBase's. PocketBase serves files directly with a token; Supabase does signed URLs with TTLs that expire cleanly. For a client with strict file-access requirements, the PocketBase setup needed a custom Go route to mimic signed URLs. Two hours of work, but worth flagging.

What it costs

For Indian solo SaaS work, the choice is between zero rupees forever (PocketBase + Oracle ARM) or twenty-five hundred rupees a month (Supabase Pro) for materially similar functionality.

When NOT to use this

Skip PocketBase if you need real Postgres features (advanced indexes, complex queries, full-text search, partitioning). PocketBase's SQLite covers basics well; for advanced workloads, Postgres is the right fit.

Skip if your team has zero Go ergonomics. PocketBase custom routes need Go; if your team only writes Python or Node, the friction is real.

Indian operator angle

For Indian solo operators and small studios bootstrapping AI SaaS, the PocketBase plus Oracle ARM stack is the clearest "build for free, scale when paying customers arrive" path I have found. Versus Supabase or Firebase, the fixed-cost difference compounds; over a year, you save Rs 25,000-30,000 that you can apply to better marketing or more Anthropic credits.

The data-residency angle matters too. Oracle Cloud's Hyderabad region keeps customer data in India by default; Supabase forces you to pick US or EU regions, which Indian compliance reviews flag for some industries (legal, fintech, healthtech).

Related

• PocketBase as SaaS backend with Claude
• Coolify deploy LLM app on Oracle ARM
• FastAPI Claude streaming endpoint