Options

Option	Type	Default	Description
output	Output	{ path: 'handlers', barrel: { type: 'named' } }	Where the generated files are written and exported
group	Group	—	Split output into per-tag or per-path folders
baseURL	string	—	Base URL prepended to every handler's request
handlers	boolean	false	Emit a handlers.ts that re-exports every handler
parser	'data' | 'faker'	'data'	Source of the response body each handler returns
include	Array<Include>	—	Keep only operations that match
exclude	Array<Exclude>	—	Skip operations that match
override	Array<Override>	—	Apply different options per pattern
resolver	Partial<ResolverMsw>	—	Customize generated names and file paths
macros	Array<Macro>	—	Rewrite AST nodes before printing

output

Where the generated handler files are written and how they are exported.

Type:	Output
Default:	{ path: 'handlers', barrel: { type: 'named' } }

output.path

Folder where the plugin writes its files. It is resolved against the global output.path on defineConfig. To write everything to one file instead, set output.mode: 'file' and give path a file name with its extension, such as 'handlers.ts'.

Type:	string
Default:	'handlers'

TIP

output.path sets where files go, output.mode sets how many. Use 'directory' (the default) for one file per operation, optionally grouped into subdirectories with the group option. Use 'file' to write everything into a single file.

output.mode

How the plugin consolidates its generated code into files.

• 'directory' (default) writes one file per operation or schema under output.path.
• 'file' writes everything into a single file. The output.path must include the file extension (for example 'handlers.ts').

Type:	'directory' | 'file'
Default:	'directory'

TIP

Pair 'directory' with the group option to organize output into per-tag or per-path subdirectories. mode: 'file' forbids group. A single-file output has nothing to group, and combining them stops the build with a KUBB_INVALID_PLUGIN_OPTIONS error.

output.barrel

Controls how the generated index.ts (barrel) file re-exports the plugin's output.

• { type: 'named' } re-exports each symbol by name. Best for tree-shaking and explicit imports.
• { type: 'all' } uses export *. Smaller barrel file, but exports everything.
• { nested: true } creates a barrel in every subdirectory, so callers can import from any depth.
• false skips the barrel entirely. The plugin's files are also excluded from the root index.ts.

Type:	{ type: 'named' | 'all', nested?: boolean } | false
Default:	{ type: 'named' }

TIP

Pick 'named' when consumers care about which symbols they import (better tree-shaking, friendlier auto-import). Pick 'all' when the file count is small and you want a one-line barrel.

'named' (default)

// src/gen/handlers/index.ts
export { getPetHandler } from './getPetHandler'
export { addPetHandler } from './addPetHandler'

'all'

// src/gen/handlers/index.ts
export * from './getPetHandler'
export * from './addPetHandler'

nested

src/gen/handlers/
├── index.ts          # re-exports ./pet and ./store
├── pet/
│   ├── index.ts      # re-exports getPetHandler, addPetHandler, ...
│   └── getPetHandler.ts
└── store/
    ├── index.ts
    └── getStoreByIdHandler.ts

false

# No index.ts is generated for this plugin.
# Its files are also excluded from the root index.ts.

output.banner

Text added to the top of every generated file. Use it for license headers, lint disables, or a @ts-nocheck directive. Pass a string for a fixed banner, or a function that builds one from a BannerMeta object. The meta carries the document info (title, description, version, baseURL) plus the per-file context filePath, baseName, isBarrel, and isAggregation, so a directive such as 'use server' can skip barrel files.

Type:	string | ((meta: BannerMeta) => string)

A static banner: '/* eslint-disable */\n// @ts-nocheck' lands at the top of each generated file:

/* eslint-disable */
// @ts-nocheck
import { http } from 'msw'

export function getPetHandler(data?: GetPetQueryResponse) {
  // ...
}

A function banner builds the text from the meta, such as banner: (meta) => \// Source: ${meta.filePath}``.

output.footer

Text added to the bottom of every generated file. It works like banner but for closing comments, such as re-enabling a lint rule. Pass a string or a function that receives the same BannerMeta and returns the text. Pair banner: '/* eslint-disable */' with footer: '/* eslint-enable */' to scope a lint disable to the generated file.

Type:	string | ((meta: BannerMeta) => string)

group

Splits generated files into subfolders by the operation's tag or URL path. Each group gets its own directory under {output.path}/{groupName}/. Without group, every file lands directly in output.path.

Type:	Group

TIP

Use group to mirror your API's domain structure (pet, store, user) in the generated code. Combine it with output.barrel: { type: 'named', nested: true } to get per-tag barrel files.

group only applies to output.mode: 'directory' (the default). It is not valid with output.mode: 'file', since a single-file output has no grouping concept.

With group: { type: 'tag' }, the generator emits one folder per tag, named after the camelCased tag:

Resulting tree

src/gen/
├── pet/
│   ├── addPetHandler.ts
│   └── getPetHandler.ts
└── store/
    ├── createStoreHandler.ts
    └── getStoreByIdHandler.ts

Pass group.name to customize the folder name. For example, a name function that appends Service to the group renames each folder to petService/.

group.type

Property used to assign each operation to a group. Required whenever group is set.

• 'tag' uses the operation's first tag.
• 'path' uses the first segment of the operation's URL, such as pet for /pet/{petId}.

An operation with no tag goes in the default group.

Type:	'tag' | 'path'

group.name

Function that turns a group key (the operation's first tag) into a folder or identifier name. The result is used as both the subdirectory name under output.path and as a suffix when naming aggregate files.

Type:	(context: { group: string }) => string
Default:	({ group }) => camelCase(group)

For type: 'path' groups, the default uses the first URL segment as-is instead of camelCasing.

baseURL

Base URL prepended to every handler's request. When omitted, the URL comes from the adapter's server URL, usually the spec's servers[0].url. Set it to point at a different environment than the spec.

Type:	string

handlers

Emits a handlers.ts file that re-exports every generated handler in operation order. Spread it into your MSW setupServer(...handlers) or setupWorker(...handlers) call.

Type:	boolean
Default:	false

gen/handlers.ts

import { getPetHandler } from './getPetHandler'
import { addPetHandler } from './addPetHandler'

export const handlers = [getPetHandler(), addPetHandler()] as const

setup.ts

import { setupServer } from 'msw/node'
import { handlers } from './gen/handlers'

export const server = setupServer(...handlers)

parser

Source of the response body each handler returns.

• 'data' (default) returns a typed empty or example payload from @kubb/plugin-ts. You fill it in from tests.
• 'faker' returns a value built by @kubb/plugin-faker. Register pluginFaker() in the plugins array. The plugin depends on Faker only when you choose this value.

Type:	'data' | 'faker'
Default:	'data'

The handler body changes with the parser. With 'data' the caller passes the payload, while 'faker' falls back to a generated mock:

'data' (default)

export function getPetHandler(data?: GetPetQueryResponse | ((info: Parameters<Parameters<typeof http.get>[1]>[0]) => Response | Promise<Response>)) {
  return http.get(`/pet/:petId`, function handler(info) {
    if (typeof data === 'function') return data(info)

    return new Response(JSON.stringify(data), {
      status: 200,
      headers: { 'Content-Type': 'application/json' },
    })
  })
}

'faker'

export function getPetHandler(data?: GetPetQueryResponse | ((info: Parameters<Parameters<typeof http.get>[1]>[0]) => Response | Promise<Response>)) {
  return http.get(`/pet/:petId`, function handler(info) {
    if (typeof data === 'function') return data(info)

    return new Response(JSON.stringify(data || getPetQueryResponse(data)), {
      status: 200,
      headers: { 'Content-Type': 'application/json' },
    })
  })
}

How you register the handler in a test depends on the value:

'data' (default)

import { setupServer } from 'msw/node'
import { getPetHandler } from './src/gen/handlers/getPetHandler'

// you pass the response payload yourself
const server = setupServer(getPetHandler({ name: 'Fluffy' }))

'faker'

import { setupServer } from 'msw/node'
import { getPetHandler } from './src/gen/handlers/getPetHandler'

// no payload needed, the handler returns generated mock data
const server = setupServer(getPetHandler())

include

Generates only the operations that match at least one entry in the list. Everything else is skipped. Each entry filters by one of:

• tag: the operation's first tag in the OpenAPI spec.
• operationId: the operation's operationId.
• path: the URL path, such as '/pet/{petId}'.
• method: the HTTP method, such as 'GET' or 'POST'.
• contentType: the request or response media type, such as 'application/json'.
• schemaName: the component schema name under #/components/schemas.

pattern accepts either a string (exact match) or a RegExp for fuzzy matches.

Type:	Array<Include>

Type definition

export type Include = {
  type: 'tag' | 'operationId' | 'path' | 'method' | 'contentType' | 'schemaName'
  pattern: string | RegExp
}

Pass include: [{ type: 'tag', pattern: 'pet' }] to keep only the pet tag. Stack entries to narrow further, such as { type: 'method', pattern: 'GET' } with { type: 'path', pattern: /^\/pet/ } for GET operations under /pet.

exclude

Skips any operation that matches at least one entry in the list. It is the opposite of include. Entries use the same type (tag, operationId, path, method, contentType, schemaName) and pattern (string or RegExp). When both are set, exclude wins.

Type:	Array<Exclude>

Type definition

export type Exclude = {
  type: 'tag' | 'operationId' | 'path' | 'method' | 'contentType' | 'schemaName'
  pattern: string | RegExp
}

Pass exclude: [{ type: 'tag', pattern: 'store' }] to drop the store tag, or stack { type: 'operationId', pattern: 'deletePet' } with { type: 'method', pattern: 'DELETE' } to skip one operation and every DELETE.

override

Applies different plugin options to operations that match a pattern. Use it for the few endpoints that need special treatment. Each entry takes the same type and pattern as include and exclude, plus an options object. That object accepts any plugin option except override, so rules cannot nest. Entries run top to bottom. The first match merges onto the plugin defaults, and later entries do not stack.

Type:	Array<Override>

Type definition

export type Override = {
  type: 'tag' | 'operationId' | 'path' | 'method' | 'contentType' | 'schemaName'
  pattern: string | RegExp
  options: Omit<Partial<Options>, 'override'>
}

For example, override: [{ type: 'tag', pattern: 'user', options: { parser: 'faker' } }] returns generated data for the user tag while the rest of the spec keeps the default parser: 'data'.

resolver

Changes how the plugin names generated files and symbols. Use it to add a prefix or suffix, or to swap the casing, without forking the plugin. Override only the methods you want to change, since anything you omit keeps its default behavior. Inside a method, this is the full resolver, so you can call this.default(name, 'function') to reuse the built-in name.

Type:	Partial<ResolverMsw> & ThisType<ResolverMsw>

TIP

Use resolver for naming and file-location tweaks. For changing the AST nodes themselves (for example stripping descriptions), use macros instead.

The default resolver names every handler with a Handler suffix and always names the aggregate export handlers.

macros

Rewrites AST nodes before they are printed to source. Use it to rename operation IDs, drop descriptions, or change schema metadata without forking the generator. Each macro callback (such as schema or operation) receives the node and a context object. Return a new node to replace it, or undefined to leave it as is. Callbacks you omit keep their default behavior. Macros run in order, so a later one sees the output of an earlier one.

Type:	Array<Macro>

TIP

Use macros to rewrite node properties before printing. For changing the names of generated symbols and files, use resolver instead.

Each entry names the macro and supplies one callback per node kind:

A macros array

import { pluginMsw } from '@kubb/plugin-msw'

pluginMsw({
  macros: [
    {
      name: 'prefix-operation-id',
      operation(node) {
        return { ...node, operationId: `api_${node.operationId}` }
      },
    },
  ],
})