StitchDB vs PlanetScale: Backend Platform or Database Service

PlanetScale vs StitchDB is a comparison of two different kinds of product. PlanetScale is a database service: fast, highly available Postgres and Vitess (MySQL) clusters on AWS, GCP and Azure, with no free plan. StitchDB is a backend platform: one Postgres database per project plus Auth, REST and GraphQL APIs, Realtime, Storage and Edge Functions, with a free plan and EU-only hosting. Pick PlanetScale when the database itself is the hard part; pick StitchDB when the work around the database is.

We make StitchDB. PlanetScale facts below are from its pricing page and plans docs, as of November 2026.

#At a glance

#What you end up building

The architectural difference shows up in the list of services you run.

With PlanetScale:

1. PlanetScale database
2. Your API server (Node, Go, Rails...) with an ORM
3. An auth provider or library
4. Object storage for files, plus resizing
5. A websocket service for live updates
6. Somewhere to run scheduled jobs

With StitchDB:

1. A StitchDB project (database, Auth, REST and GraphQL, Storage, Realtime, Edge Functions, cron)
2. Optionally your own server, for logic that doesn't fit in SQL or functions

If you already have items 2 to 6, PlanetScale slots in with no wasted features. If you don't, StitchDB removes most of them.

#Performance and availability

This is PlanetScale's strength. Production Postgres clusters run a primary with two replicas in separate availability zones, replicas can be promoted on failover, and Metal plans put the database on local NVMe drives. Vitess adds horizontal sharding for MySQL workloads, and Neki brings sharding to Postgres.

StitchDB is built differently. Each project's database runs on a shared server by default, with plan limits on connections and request rates per project. You can add read replicas ($0.01 per replica-hour) on Pro and up, and Scale and Enterprise can turn on automatic replicas that are added under load. Per the compute docs, Scale and Enterprise projects can also move to a dedicated server, from 4 GB of RAM up to 384 GB or more, with no change to connection strings. High availability with automatic failover is not listed as a plan feature.

If you need guaranteed multi-AZ failover today, PlanetScale is the safer choice.

#Pricing at three sizes

All prices as of November 2026.

Side project. PlanetScale: $5 a month for a single-node PS-5 (no replicas). StitchDB: $0 on Free, with 500 MB of database and 24 hours of point-in-time recovery.

Small production app. PlanetScale: $15 a month for a PS-5 HA cluster, plus your auth and storage providers. StitchDB: $15.99 a month on Pro with 10 GB of database, 7 days of PITR, 100,000 MAU of auth, 100 GB of files and 1 TB of egress included.

Growing app needing more compute. PlanetScale scales through larger PS sizes or Metal (M-10 HA from $50). StitchDB moves to Scale ($49) and, if needed, a dedicated server billed hourly from prepaid credit.

#Branching

The two take different approaches.

PlanetScale's Vitess branches come with deploy requests: you propose a schema change on a branch, review it, and deploy it as a non-blocking online change, with the option to revert. For Postgres, development branches are single-instance databases at $5 a month each.

StitchDB preview branches are separate projects built from production's migrations, with their own URL, keys, Auth users and storage. You can copy production data with personal data masked, where masking runs inside production's own query:

stitchdb branches create feature/checkout --with-data --wait
stitchdb branches switch feature/checkout
stitchdb db push
stitchdb branches merge feature/checkout --dry-run
stitchdb branches merge feature/checkout --delete

Merging applies the branch's new migrations to production. Data isn't merged. Pro includes 3 branches; Free has none.

#Developer experience

PlanetScale is a database you connect to with any Postgres or MySQL driver. Its query insights help with slow queries, and Traffic Control is available for Postgres.

StitchDB lets the client talk to the database directly through the generated API, with Row Level Security doing authorization:

import { createClient } from 'stitchdb-js'

const stitchdb = createClient('https://<project-ref>.p.stitchdb.com', '<anon-key>')

await stitchdb.auth.signInWithPassword({ email, password })

const { data } = await stitchdb
  .from('projects')
  .select('id, name, tasks(id, title, done)')
  .order('name')

You can still connect with Prisma, Drizzle, psql or any Postgres client on ports 5432 and 6543 (Connect).

#Who should pick which

Pick PlanetScale if:

• you run MySQL at scale and depend on Vitess,
• you need multi-AZ HA on day one,
• your users are in the US or Asia, or you need a specific cloud,
• you already have an API layer, auth and storage.

Pick StitchDB if:

• you want one project to cover database, auth, files and APIs,
• you want a free tier for prototypes and staging,
• your users are mainly in Europe or you need EU hosting,
• point-in-time recovery on a small budget matters.

Also consider Neon vs PlanetScale if you are only choosing a database.

#Moving from PlanetScale

From PlanetScale Postgres, use pg_dump or StitchDB's Import from Postgres, which runs a live copy when the source allows logical replication. From PlanetScale Vitess (MySQL), you first convert to Postgres: see Migrate from PlanetScale (MySQL) to Postgres.

#Connections from serverless functions

Both platforms expect serverless code, and both solve the connection problem, in different ways.

On PlanetScale Postgres, you connect through PgBouncer on port 6432 for application traffic and directly on port 5432 for migrations and sessions that need state. TLS is required on both.

On StitchDB, every connection goes through the pooler: port 5432 in session mode, port 6543 in transaction mode, TLS required. For runtimes that can't open TCP at all (Cloudflare Workers, Vercel Edge, Deno Deploy), there is an HTTP and WebSocket driver that runs queries through a pool on the project's server:

import { createSql } from 'stitchdb-js/sql'

const sql = createSql(env.STITCHDB_SQL_URL)

const [, [order]] = await sql.transaction([
  sql`update stock set left = left - 1 where sku = ${sku}`,
  sql`insert into orders (sku, buyer) values (${sku}, ${userId}) returning *`,
])

The SQL key runs as the project owner, so it belongs in server or worker code only (serverless driver). And for code on users' devices, StitchDB's REST API with the anon key and Row Level Security avoids database connections entirely.

#FAQ

#Is PlanetScale free?

No. As of November 2026 PlanetScale has no free plan; Postgres starts at $5 a month for a single node.

#Is StitchDB faster than PlanetScale?

We don't publish benchmarks. PlanetScale's Metal plans with local NVMe and HA clusters are built for high performance; StitchDB focuses on bundling the backend around Postgres.

#Does StitchDB support MySQL?

No. StitchDB is Postgres only.

#Can I use PlanetScale with StitchDB Auth?

Not directly: StitchDB Auth issues JWTs for its own database and policies. You could verify StitchDB JWTs in your own API in front of PlanetScale, but at that point you are running two platforms.

Want the backend included? Start free on StitchDB or compare plans.