Class: Request
HTTP client with interceptors, timeout, retry, cache, deduplication, and download progress.
Basic usage
const api = new Request('https://api.example.com')
// Directly awaitable
const user = await api.get<User>('/users/1')
// Lifecycle callbacks
await api.get('/users', {
onSuccess: setUsers,
onError: toast.error,
onFinally: () => setLoading(false),
})
Composable best practice
Prefer small focused instances over one instance with all interceptors:
// ── Building blocks (pure functions, reusable) ──
const addAuth: RequestInterceptor = (config) => ({
...config,
headers: { ...config.headers, Authorization: `Bearer ${getToken()}` }
})
const addLang: RequestInterceptor = (config) => ({
...config,
headers: { ...config.headers, 'Accept-Language': 'zh-CN' }
})
// ── Compose: each instance handles one concern ──
const authApi = new Request('https://api.example.com')
authApi.useRequestInterceptor(addAuth)
authApi.useRequestInterceptor(addLang)
const publicApi = new Request('https://open.api.com')
// ── Or use per-request interceptors ──
api.get('/users', {}, {
requestInterceptors: [addAuth],
onSuccess: setUsers,
})
Constructors
Constructor
new Request(baseUrl?): Request
Parameters
baseUrl?
string = ''
Returns
Request
Methods
clearCache()
clearCache(): void
Clear the in-memory response cache.
Returns
void
delete()
delete<T>(url, params?, config?): RequestHandle<T>
Type Parameters
T
T = unknown
Parameters
url
string
params?
Record<PropertyKey, any>
config?
RequestConfig = {}
Returns
RequestHandle<T>
get()
get<T>(url, params?, config?): RequestHandle<T>
Type Parameters
T
T = unknown
Parameters
url
string
params?
Record<PropertyKey, any>
config?
RequestConfig = {}
Returns
RequestHandle<T>
patch()
patch<T>(url, data?, config?): RequestHandle<T>
Type Parameters
T
T = unknown
Parameters
url
string
data?
unknown
config?
RequestConfig = {}
Returns
RequestHandle<T>
post()
post<T>(url, data?, config?): RequestHandle<T>
Type Parameters
T
T = unknown
Parameters
url
string
data?
unknown
config?
RequestConfig = {}
Returns
RequestHandle<T>
put()
put<T>(url, data?, config?): RequestHandle<T>
Type Parameters
T
T = unknown
Parameters
url
string
data?
unknown
config?
RequestConfig = {}
Returns
RequestHandle<T>
request()
request<T>(url, config?): RequestHandle<T>
Type Parameters
T
T = unknown
Parameters
url
string
config?
RequestConfig = {}
Returns
RequestHandle<T>
sse()
sse<T>(url, config?): RequestHandle<T>
Create an SSE (Server-Sent Events) connection.
Uses Fetch API under the hood, so it supports:
- Custom headers (unlike native EventSource)
- POST requests
- All HTTP methods
Type Parameters
T
T = unknown
Parameters
url
string
config?
Omit<RequestConfig, "sse"> & SSECallbacks = {}
Returns
RequestHandle<T>
Example
const api = new Request('https://api.example.com')
// GET SSE (default)
const handle = api.sse('/events', {
onMessage: (event) => console.log(event.data),
onError: (err) => console.error(err),
})
// POST SSE with data
const handle = api.sse('/chat/completions', {
method: 'POST',
data: { prompt: 'Hello' },
onMessage: (event) => {
const parsed = JSON.parse(event.data)
console.log(parsed)
},
})
// Cancel the connection
handle.abort()
```n
***
### useErrorInterceptor()
> **useErrorInterceptor**(`interceptor`): () => `void`
#### Parameters
##### interceptor
[`ErrorInterceptor`](../types/ErrorInterceptor.md)
#### Returns
() => `void`
***
### useRequestInterceptor()
> **useRequestInterceptor**(`interceptor`): () => `void`
#### Parameters
##### interceptor
[`RequestInterceptor`](../types/RequestInterceptor.md)
#### Returns
() => `void`
***
### useResponseInterceptor()
> **useResponseInterceptor**\<`T`, `U`\>(`interceptor`): () => `void`
#### Type Parameters
##### T
`T` = `unknown`
##### U
`U` = `unknown`
#### Parameters
##### interceptor
[`ResponseInterceptor`](../types/ResponseInterceptor.md)\<`T`, `U`\>
#### Returns
() => `void`
