Supabase vs. Raw PostgreSQL vs. MySQL: Which Database for Your Automation Stack?

When you need a database for automation

Most automation work runs on CRMs. But sometimes you need more:

• Custom data the CRM can't hold well
• High-volume logs or events
• Cross-CRM sync with a source-of-truth
• Analytics and reporting beyond CRM dashboards
• Multi-tenant data for SaaS-like offerings

That's when you reach for a real database. The three common choices for small/mid operations: Supabase, raw PostgreSQL (self-hosted or on a managed provider), and MySQL.

---

TL;DR

• Supabase: Managed Postgres + REST API + auth + real-time. Start here if you want productivity over control.
• Raw PostgreSQL: Self-hosted or on AWS RDS/DigitalOcean. Choose when you need full control or your stack already uses Postgres.
• MySQL: Legacy-friendly, widely hosted cheaply. Pick when an integration requires it or when cPanel hosting is your only option.

For most automation stacks, Supabase wins on speed-to-productive.

---

Supabase deep dive

What it is

A managed PostgreSQL platform with layers on top: auto-generated REST + GraphQL API, auth (email/password, OAuth, magic links), realtime subscriptions, storage (S3-like), and edge functions.

Pricing (April 2026)

• Free: 500 MB database, 1 GB storage, 50,000 monthly active users
• Pro: $25/month per project — 8 GB database, 100 GB storage, 100k MAUs, daily backups
• Team: $599/month — higher limits, SLA, support
• Enterprise: custom

For most small automation stacks, Free or Pro is sufficient.

What's good

• Instant API: Define a table, get a REST API automatically. No backend code needed.
• Row-Level Security (RLS): Secure per-user data access at the database level.
• Real-time subscriptions: Websocket updates when rows change. Useful for dashboards, live notifications.
• Auth built-in: Don't build sign-up/login yourself.
• Storage with S3 API: File uploads integrated with database.
• Edge functions: Run TypeScript serverless functions.
• Dashboard: GUI for queries, table editing, and user management.

What's not so good

• Cold starts on free tier (dormant projects pause after 7 days of inactivity).
• Vendor lock-in risk if you use Supabase-specific features (auth, edge functions, RLS) — moving away is harder.
• Pricing jumps at Team tier — $25 to $599 is a big gap.
• No built-in job scheduler (you need external cron or edge functions on schedule).

When Supabase wins

• You want a database + API without writing backend code
• You're building a custom app on top of your CRM data
• You need real-time features (live dashboards, notifications)
• You want user authentication without rolling your own

---

Raw PostgreSQL deep dive

What it is

The PostgreSQL database server itself. Run it on a VPS (DigitalOcean, Hetzner, AWS EC2), a managed service (AWS RDS, DigitalOcean Managed Postgres), or your own hardware.

Costs

• Self-hosted on VPS: $5-$20/month for a small instance
• AWS RDS (small): ~$15-$30/month + data transfer
• DigitalOcean Managed: $15/month for 1 GB RAM / 10 GB disk
• Hetzner: $5-$10/month for VPS to self-host

What's good

• No vendor lock-in. Standard Postgres. Works anywhere.
• Full control. Any Postgres feature, any extension, any config.
• Cheap at scale. $20/month manages lots of data and queries.
• Mature. Decades of reliability, extensive tooling.

What's not so good

• You build the API. No auto-generated REST/GraphQL. Need to write a backend or use tools like PostgREST, Hasura.
• No auth built-in. You build login/signup yourself.
• Backup is your job if self-hosted. Managed services handle it.
• No GUI. Use pgAdmin, DBeaver, TablePlus.

When raw Postgres wins

• Your stack is developer-heavy and you have backend engineers
• You need full control for compliance or custom configuration
• You're running on existing AWS infrastructure
• You're building long-term and want to avoid vendor dependencies

---

MySQL deep dive

What it is

The other major open-source database. Widely deployed, widely hosted, especially on cheap shared hosting (cPanel, WordPress hosts).

Costs

• Shared hosting: often included with WordPress/LAMP hosting ($5-$15/month)
• Managed MySQL (AWS RDS, DigitalOcean): $15+/month
• Self-hosted: similar to Postgres

What's good

• Ubiquity. Every hosting provider supports it. WordPress uses it.
• Familiar. Many devs learned MySQL first.
• Strong for simple read-heavy workloads. Fast for typical CRUD.

What's not so good

• Weaker for complex queries than Postgres. Postgres has better analytics, JSON support, window functions, CTE optimization.
• JSON support is okay, not great. Postgres's JSONB is better.
• Less feature-rich extensions. Postgres has PostGIS, pg_trgm, extensions Postgres is known for.
• MariaDB compatibility confusion. Some hosts use MariaDB (MySQL fork), features drift.

When MySQL wins

• Integration requires it (WordPress plugin, legacy app)
• Your hosting only supports MySQL (shared cPanel)
• Your team is MySQL-first and doesn't want to switch

For greenfield automation stacks, MySQL rarely wins in 2026. It's fine but Postgres is generally a better default.

---

Feature comparison

| Feature | Supabase | Raw Postgres | MySQL | |---------|----------|--------------|-------| | Base cost (small) | $0-$25/mo | $5-$20/mo | $5-$15/mo | | Auto API | Yes (REST + GraphQL) | No (build yourself or use PostgREST) | No | | Built-in auth | Yes | No | No | | Realtime updates | Yes | No (without extensions) | No | | Setup time | <10 min | 1-2 hours | 1-2 hours | | Vendor lock-in | Moderate | None | None | | JSON support | Excellent (Postgres JSONB) | Excellent | Decent | | Full-text search | Yes (Postgres) | Yes | Yes (limited) | | Community | Growing | Huge | Huge |

---

Real automation stack examples

Example 1: Small consultancy (Supabase)

Use case: custom CRM-like contact database that syncs with GoHighLevel, plus a custom dashboard showing metrics the GHL dashboard doesn't.

• Supabase free tier (500 MB database)
• Tables: contacts, events, metrics_daily
• Next.js frontend using Supabase auto-generated API
• GoHighLevel webhooks write to events table via Make
• Daily scheduled edge function aggregates events into metrics_daily
• Cost: $0-$25/month

Example 2: Agency managing multi-tenant (Supabase Pro)

Use case: agency running 15 client campaigns, each with its own CRM data and dashboard.

• Supabase Pro ($25/month)
• Row-Level Security: each client's user only sees their own data
• Real-time dashboard: new leads appear immediately
• Storage: client-facing file uploads
• Cost: $25-$75/month

Example 3: Developer-led team (Raw Postgres)

Use case: software company building a custom product with a backend team.

• AWS RDS Postgres (~$50/month)
• Custom Node.js/Python API
• Auth handled by company's existing auth service
• PostgREST for rapid prototyping
• Cost: $50-$200/month depending on scale

---

Migration paths

From MySQL to Postgres:

• Tool: pgloader (open source). Good compatibility, handles most cases.
• Manual: export CSV from MySQL, import to Postgres. Simpler for small datasets.

From Postgres to Supabase:

• Export Postgres dump, import into Supabase via SQL editor or CLI.
• Easy because Supabase IS Postgres.

From Supabase to raw Postgres:

• Export via Supabase CLI or pg_dump.
• Drop features (auth, storage, edge functions) that don't have direct equivalents.

---

What I recommend

Default choice for new automation stacks in 2026: Supabase.

Reasons:

• Fast to start (database + API in minutes)
• Scales well up to mid-size workloads
• Postgres under the hood means no lock-in on data (you can export to raw Postgres)
• Free tier is generous enough for prototyping

Pick raw Postgres when:

• You have a backend engineering team that wants control
• You're on existing AWS/GCP infrastructure
• You need extensions Supabase doesn't expose

Pick MySQL when:

• External requirement (WordPress, legacy app)
• Hosting constraint (cheap cPanel)
• Team expertise heavily MySQL

---

Sources

Pricing and feature data from supabase.com/pricing, postgresql.org/support (links to managed providers), mysql.com, and AWS RDS/DigitalOcean pricing pages as of April 2026. Postgres vs. MySQL feature comparisons are widely documented across database engineering literature and benchmarking reports.

Choosing a database for your automation stack? Let's talk — I can help you decide based on your specific use case.