TypeSQL vs Prisma ORM

Prisma ORM describes your database in a Prisma schema (schema.prisma). prisma generate turns it into a typed client, and prisma migrate creates and applies migrations.

The same query

With Prisma ORM:

const users = await prisma.user.findMany({
	where: { createdAt: { gt: new Date("2026-01-01") } },
	select: { id: true, email: true, posts: { select: { title: true } } },
});

With TypeSQL:

const rows = await sql(
	`SELECT u.id, u.email, p.title
   FROM users u
   LEFT JOIN posts p ON p.author_id = u.id
   WHERE u.created_at > ${sql.param(new Date("2026-01-01"))}`,
).execute();

Prisma returns nested objects (each user with a posts array). TypeSQL returns the rows your SQL produces; nest them with JSON aggregation when you want that shape.

How they differ

	Prisma ORM	TypeSQL
You write	a Prisma schema and client calls	a schema type and SQL
Types come from	the client prisma generate writes	the TypeScript compiler, reading your schema type
Extra step	prisma generate after each schema change	none
Raw SQL	$queryRaw, typed by the type you write, or TypedSQL	checked and typed from the SQL
Migrations	prisma migrate generates them from the schema	you write the DDL, and the schema type follows it
Databases	PostgreSQL, MySQL, SQLite, SQL Server, MongoDB, CockroachDB	PostgreSQL and SQLite; MySQL is planned

Choose Prisma ORM when

• You want relations, nested reads and nested writes without writing joins.
• You want migrations generated from schema changes.
• You use a database TypeSQL doesn't support yet.

Choose TypeSQL when

• You'd rather write the SQL itself, including joins, CTEs, window functions and anything else your database supports.
• You want no generate step between changing a query and seeing its types.
• You want queries you can copy into psql or a SQL log and run as they are.

Using them together

@typesql/prisma reads your Prisma schema, through the typesql-prisma generator or Prisma Client's Prisma.TypeMap, and runs queries on Prisma Client or a transaction client. Keep Prisma for the schema, migrations and simple reads, and write the rest as SQL. See From Kysely, Drizzle or Prisma.