How to use
- Copy the installation command and review the source package before installing it in your own agent environment.
- Check the source document for required tools, credentials and supported hosts.
- Describe your input and desired deliverable; review the output against the source material.
npx skills add https://github.com/TerminalSkills/skills --skill kyselyBefore you start
- Installation does not include a model subscription, API allowance or permission to access your files.
- Repository stars and installation counts describe activity, not an independent quality or security test.
Rendered from the source repository. Headings, examples, code, tables, links, and referenced images are preserved.
Kysely
Overview
Kysely is a type-safe TypeScript SQL query builder. Unlike ORMs, it doesn't abstract away SQL — it gives you full SQL power with TypeScript autocompletion and type checking. Every query is validated at compile time. Kysely generates SQL and hands it to your driver, nothing more.
Kysely 0.29 is ESM-only and needs Node.js 22 or newer (engines in its package.json) and TypeScript 5.4+. Install the driver yourself: npm install kysely pg (PostgreSQL), mysql2, better-sqlite3, or tedious tarn (MSSQL). Dialects ship in the package: PostgresDialect, MysqlDialect, SqliteDialect, MssqlDialect, PGliteDialect.
Instructions
Step 1: Define Types
// db/types.ts — Database type definitions
import type { Generated, Insertable, Selectable, Updateable } from 'kysely'
export interface Database {
users: UsersTable
posts: PostsTable
}
interface UsersTable {
id: Generated<number>
name: string
email: string
created_at: Generated<Date> // node-postgres returns timestamps as Date
}
interface PostsTable {
id: Generated<number>
title: string
body: string
author_id: number
published: boolean
created_at: Generated<Date>
}
// Helper types: Generated columns are optional on insert and update.
// Use ColumnType<Select, Insert, Update> for a column that differs per operation,
// e.g. ColumnType<Date, string | undefined, never> for a created_at nobody may change.
type NewUser = Insertable<UsersTable>
type UserUpdate = Updateable<UsersTable>
type User = Selectable<UsersTable>To generate these interfaces from a live database instead of writing them, install kysely-codegen (0.20) as a dev dependency and run npx kysely-codegen --out-file ./src/db/db.d.ts with DATABASE_URL set; it reads the schema and picks the dialect from the URL.
Step 2: Queries
// db/queries.ts — Type-safe SQL queries
import { Kysely, PostgresDialect, sql } from 'kysely'
import { Pool } from 'pg'
import type { Database } from './types'
const db = new Kysely<Database>({
dialect: new PostgresDialect({ pool: new Pool({ connectionString: process.env.DATABASE_URL }) }),
})
// Select with joins — fully typed result
const postsWithAuthor = await db
.selectFrom('posts')
.innerJoin('users', 'users.id', 'posts.author_id')
.select(['posts.id', 'posts.title', 'users.name as author_name'])
.where('posts.published', '=', true)
.orderBy('posts.created_at', 'desc')
.limit(20)
.execute()
// postsWithAuthor is { id: number, title: string, author_name: string }[]
// Subquery
const activeAuthors = await db
.selectFrom('users')
.select(['users.name', 'users.email'])
.where('users.id', 'in',
db.selectFrom('posts')
.select('posts.author_id')
.where('posts.published', '=', true)
.groupBy('posts.author_id')
)
.execute()
// Insert
const newUser = await db
.insertInto('users')
.values({ name: 'Alice Moreau', email: 'alice.moreau@northwind.io' })
.returningAll()
.executeTakeFirstOrThrow()
// Transaction
await db.transaction().execute(async (trx) => {
const user = await trx.insertInto('users')
.values({ name: 'Bob Okafor', email: 'bob.okafor@northwind.io' })
.returningAll()
.executeTakeFirstOrThrow()
await trx.insertInto('posts')
.values({ title: 'First Post', body: 'Hello!', author_id: user.id, published: true })
.execute()
}) // commits when the callback resolves, rolls back when it throws
// Update and delete: always add a where(); without one every row changes
await db.updateTable('users').set({ name: 'Alice M.' }).where('id', '=', newUser.id).execute()
await db.deleteFrom('posts').where('author_id', '=', newUser.id).execute()
// Count: PostgreSQL returns bigint as a string, so convert it
const { total } = await db.selectFrom('posts')
.select((eb) => eb.fn.countAll<string>().as('total'))
.executeTakeFirstOrThrow()
console.log(Number(total))Step 3: Migrations
Put one file per migration in a folder (001_create_users.ts, 002_...; they run in name order) and run them with the Migrator. In 0.29 import it from kysely/migration; the root kysely export of Migrator and FileMigrationProvider is a type error (checked).
// migrate.ts — run with: npx tsx migrate.ts
import * as path from 'node:path'
import { promises as fs } from 'node:fs'
import { Migrator, FileMigrationProvider } from 'kysely/migration'
const migrator = new Migrator({
db, // the Kysely instance from Step 2
provider: new FileMigrationProvider({ fs, path, migrationFolder: path.join(process.cwd(), 'migrations') }),
})
const { error, results } = await migrator.migrateToLatest()
results?.forEach((r) => console.log(`${r.status}: ${r.migrationName}`))
if (error) { console.error(error); process.exit(1) }
await db.destroy()// migrations/001_create_users.ts
import { Kysely, sql } from 'kysely'
export async function up(db: Kysely<any>) {
await db.schema
.createTable('users')
.addColumn('id', 'serial', (col) => col.primaryKey())
.addColumn('name', 'varchar(255)', (col) => col.notNull())
.addColumn('email', 'varchar(255)', (col) => col.notNull().unique())
.addColumn('created_at', 'timestamp', (col) => col.defaultTo(sql`now()`).notNull())
.execute()
}
export async function down(db: Kysely<any>) {
await db.schema.dropTable('users').execute()
}Examples
Example 1: Paginated posts with author names
User prompt: "List the 20 newest published posts with the author's name, using Kysely on our Postgres database."
The agent adds PostsTable and UsersTable to the Database interface, then runs the join from Step 2. The inferred type is { id: number; title: string; author_name: string }[]; a typo such as 'posts.titel' fails tsc before anything runs. For page 3 it adds .limit(20).offset(40).
Example 2: Add a table and apply it
User prompt: "Add a comments table with a foreign key to posts and run the migration."
The agent writes migrations/002_create_comments.ts with db.schema.createTable('comments'), an integer post_id column using .references('posts.id').onDelete('cascade'), a matching down that drops it, then runs npx tsx migrate.ts. The output is Success: 002_create_comments. It then adds CommentsTable to Database, or regenerates the types with kysely-codegen.
Guidelines
- Kysely is a query builder, not an ORM — no relations, no lazy loading, no magic. Just SQL with types.
- Use Kysely when you want SQL control with type safety. Use Drizzle or Prisma when you want ORM features.
- Create one
Kyselyinstance per process and share it; calldb.destroy()on shutdown. Never build SQL by string concatenation: use thesqltemplate tag, which keeps interpolated values as bound parameters. - Serverless databases work through their own drivers or community dialects; check that the driver supports transactions before relying on them.
- Column names in
.select(),.where()and.orderBy()are type-checked, but the real schema is not: keepDatabasein sync with migrations (kysely-codegen helps). - The
Insertable/Updateabletypes automatically makeGeneratedfields optional.

