createRequest()

createRequest(defaults: RequestDefaults | null): RequestInstance creates an independent request instance (mirroring axios.create). The instance has its own RequestDefaults and interceptor registry, fully isolated from the default instance.

import { createRequest } from '@/uni_modules/ux-request'

const api = createRequest({
	baseURL: 'https://api.example.com',
	timeout: 10000,
	header: null,
	validateStatus: null
})

api.get<UTSJSONObject>('/users/1', null).then((res) => {
	console.log(res.data)
})

Why Multiple Instances

The default instance (top-level APIs such as request()) is a module-level singleton; the defaults modified by configureRequest and the interceptors registered by useRequestInterceptor all live on it. The following scenarios need mutually isolated instances:

• Multiple backends: the business API and third-party APIs each have their own baseURL, timeout, and auth header;
• Interceptors that don't pollute each other: a token header injected on instance A should not appear on instance B's requests (registries are isolated per instance, see the example below);
• Local environment reproduction: e.g. reproducing the ERR_CONFIG error ("relative path without baseURL") on an isolated instance, unaffected by the global baseURL.

Parameters

createRequest(defaults: RequestDefaults | null)

Parameter	Type	Description
defaults	RequestDefaults | null	Instance default configuration. null means all built-in defaults apply (no baseURL / timeout 60000 / empty header / 2xx validateStatus)

Return Value

Returns a RequestInstance — an instance object structurally identical to the default instance; see Request Instance for the member list. Instance methods provide no generics; data is returned as any (ResponseAny), cast as needed on the business side.

Isolated interceptor registries

Each instance maintains its own request/response interceptor registries: an interceptor registered via useRequestInterceptor on instance A only applies to A's requests, and ejectInterceptor only looks up A's registry. For the default instance's interceptors see Interceptor API.

Examples

Two backends, isolated

import { createRequest } from '@/uni_modules/ux-request'

// Business API: long timeout + auth header
const authHeader: UTSJSONObject = {}
authHeader['Authorization'] = 'Bearer <token>'

const api = createRequest({
	baseURL: 'https://api.example.com',
	timeout: 15000,
	header: authHeader,
	validateStatus: null
})

// Object storage: short timeout, no auth
const oss = createRequest({ baseURL: 'https://oss.example.com', timeout: 5000, header: null, validateStatus: null })

api.get<UTSJSONObject>('/me', null)
oss.get<UTSJSONObject>('/bucket/a.png', null)

Instance-level interceptor isolation

import { createRequest } from '@/uni_modules/ux-request'

const a = createRequest({ baseURL: 'https://a.test', timeout: 10000, header: null, validateStatus: null })
const b = createRequest({ baseURL: 'https://b.test', timeout: 10000, header: null, validateStatus: null })

// Only applies to a
const rid = a.useRequestInterceptor((config) => {
	const h = config.header
	if (h != null) {
		h['x-only-a'] = 'A'
	}
	return config
})

a.request({ url: '/x' }).then((resA) => {
	b.request({ url: '/x' }).then((resB) => {
		// resA.config carries x-only-a, resB.config does not
		a.ejectInterceptor(rid)
	})
})

Related APIs

• Request Instance
• RequestInstance
• RequestDefaults
• configureRequest()
• Multiple Instances and Global Defaults