How to Migrate From Supabase — Complete Step-by-Step Guide

March 31, 2026 · 12 min read · Updated regularly

This guide walks through migrating from Supabase to another platform — step by step, with actual commands and scripts. We'll cover database export, auth migration, storage transfer, RLS policy translation, and zero-downtime cutover.

Target audience: developers who've decided to move off Supabase (whether due to the India block, pricing, or wanting more control).

Step 0: Decide Your Target

Before migrating, know where you're going:

• Neon — Serverless Postgres. Easiest migration from Supabase because it's still Postgres. Your SQL, schemas, and most queries work as-is.
• Self-hosted Supabase — Same stack, your server. Zero code changes but you manage infrastructure.
• Firebase — Google's BaaS. Full rewrite of your data layer (NoSQL vs SQL).
• Custom stack — Postgres (Neon/Railway) + Clerk/Auth0 + S3/R2. Most work upfront, most flexibility long-term.

Not sure? Take our free migration assessment — it recommends the best target based on your project.

Step 1: Export Your Database

Supabase gives you full Postgres access, so you can use standard tools:

Get your connection string

Go to your Supabase dashboard → Settings → Database → Connection string. Copy the URI format.

Export schema + data

# Export everything (schema + data)
pg_dump "postgresql://postgres:[PASSWORD]@db.[PROJECT_REF].supabase.co:5432/postgres" \
  --no-owner \
  --no-privileges \
  -F c \
  -f supabase_backup.dump

# Or for plain SQL (easier to inspect/edit):
pg_dump "postgresql://postgres:[PASSWORD]@db.[PROJECT_REF].supabase.co:5432/postgres" \
  --no-owner \
  --no-privileges \
  --schema=public \
  -f supabase_backup.sql

Export auth users separately

# Export auth.users table
pg_dump "postgresql://postgres:[PASSWORD]@db.[PROJECT_REF].supabase.co:5432/postgres" \
  --no-owner \
  --schema=auth \
  --table=auth.users \
  -f auth_users.sql

Step 2: Import to Your New Database

For Neon / Railway / Any Postgres

# Restore from dump
pg_restore -d "postgresql://[USER]:[PASSWORD]@[HOST]/[DB]" \
  --no-owner \
  --no-privileges \
  supabase_backup.dump

# Or from SQL file
psql "postgresql://[USER]:[PASSWORD]@[HOST]/[DB]" \
  -f supabase_backup.sql

Verify your tables are there:

psql "postgresql://[USER]:[PASSWORD]@[HOST]/[DB]" \
  -c "\dt"

For Firebase (Firestore)

There's no direct SQL → Firestore migration. You need to:

1. Export your Supabase data as JSON (use psql with \copy or a script)
2. Write a migration script that maps your relational data to Firestore's document model
3. Denormalize joins into nested documents or subcollections
4. Import using Firebase Admin SDK

This is the hardest part of a Firebase migration. Budget 1-3 days depending on schema complexity.

Step 3: Migrate Authentication

Supabase uses GoTrue for auth. Your user data is in the auth.users table.

To Clerk

// Export users from Supabase auth.users
// Then import via Clerk Backend API:
const clerk = require('@clerk/clerk-sdk-node');

for (const user of supabaseUsers) {
  await clerk.users.createUser({
    emailAddress: [user.email],
    password: user.encrypted_password, // Clerk supports bcrypt import
    firstName: user.raw_user_meta_data?.first_name,
    lastName: user.raw_user_meta_data?.last_name,
  });
}

To Firebase Auth

// Use Firebase Admin SDK bulk import
const admin = require('firebase-admin');

const users = supabaseUsers.map(u => ({
  uid: u.id,
  email: u.email,
  passwordHash: Buffer.from(u.encrypted_password),
  passwordSalt: Buffer.from(''), // Supabase uses bcrypt (salt embedded)
}));

// Firebase supports bcrypt password hashes
await admin.auth().importUsers(users, {
  hash: { algorithm: 'BCRYPT' }
});

Step 4: Migrate Storage

Supabase Storage is built on S3-compatible APIs. Your files are in buckets.

Download all files

// Using Supabase JS client
const { data: buckets } = await supabase.storage.listBuckets();

for (const bucket of buckets) {
  const { data: files } = await supabase.storage
    .from(bucket.name)
    .list('', { limit: 1000 });
  
  for (const file of files) {
    const { data } = await supabase.storage
      .from(bucket.name)
      .download(file.name);
    
    // Save to disk or upload to new storage
    fs.writeFileSync(`./backup/${bucket.name}/${file.name}`, Buffer.from(await data.arrayBuffer()));
  }
}

Upload to new storage

For Cloudflare R2 (S3-compatible, free egress):

// Using AWS SDK (R2 is S3-compatible)
const { S3Client, PutObjectCommand } = require('@aws-sdk/client-s3');

const r2 = new S3Client({
  region: 'auto',
  endpoint: 'https://[ACCOUNT_ID].r2.cloudflarestorage.com',
  credentials: { accessKeyId: R2_KEY, secretAccessKey: R2_SECRET }
});

// Upload each file
await r2.send(new PutObjectCommand({
  Bucket: 'my-bucket',
  Key: file.name,
  Body: fileBuffer,
  ContentType: file.contentType
}));

Step 5: Translate RLS Policies

Supabase's Row-Level Security policies run in Postgres. If your new database is Postgres (Neon, self-hosted), your RLS policies migrate as-is.

If you're moving to a non-Postgres platform, you need to translate RLS to application-level security:

// Supabase RLS policy (Postgres):
// CREATE POLICY "Users can only see their own data"
//   ON public.profiles FOR SELECT
//   USING (auth.uid() = user_id);

// Translated to application code (e.g., Next.js API route):
async function getProfile(userId) {
  // Verify the requesting user matches the profile owner
  const session = await getSession();
  if (session.userId !== userId) {
    throw new Error('Unauthorized');
  }
  return db.query('SELECT * FROM profiles WHERE user_id = $1', [userId]);
}

Step 6: Update Your App

Replace the Supabase client with your new stack's client. This is where the effort varies dramatically:

If you moved to Neon (still Postgres)

// Before (Supabase)
import { createClient } from '@supabase/supabase-js'
const supabase = createClient(URL, KEY)
const { data } = await supabase.from('users').select('*')

// After (Neon + Drizzle/Prisma)
import { neon } from '@neondatabase/serverless'
const sql = neon(DATABASE_URL)
const data = await sql('SELECT * FROM users')

If you moved to Firebase

// Before (Supabase)
const { data } = await supabase.from('users').select('*').eq('id', userId)

// After (Firebase)
const doc = await getDoc(doc(db, 'users', userId))
const data = doc.data()

Every query needs rewriting. Joins become subcollection queries or denormalized reads.

Step 7: Zero-Downtime Cutover

1. Set up the new stack in parallel — don't touch your existing Supabase config yet
2. Run a final data sync — export + import the latest data from Supabase
3. Switch your app's config — update environment variables to point to the new database/auth/storage
4. Deploy — one deployment switches everything
5. Monitor for 24 hours — watch for auth failures, missing data, broken queries
6. Keep Supabase running for 1 week — as a rollback option. Then decommission.

Migration Checklist

• ☐ Database schema exported and imported
• ☐ Data verified in new database (row counts match)
• ☐ Auth users migrated (test login with existing credentials)
• ☐ Storage files transferred (verify file counts and access)
• ☐ RLS policies translated to new platform or app-level security
• ☐ All Supabase client calls replaced
• ☐ Environment variables updated
• ☐ Staging test passed end-to-end
• ☐ Final data sync completed
• ☐ Production deployed and monitoring active

Related

• Supabase Blocked in India? Here's How to Fix It in 60 Seconds
• 7 Supabase Alternatives for Indian Developers (2026)
• SupaRoute — Managed Supabase Proxy
• SupaMigrate — Migration Assessment + Services