fetchja

KO | EN

gitlite — search

#javascript #python #hacktoberfest #react #ai #typescript #llm #go #golang #android #machine-learning #rust #deep-learning #linux

(★ 12)

A tiny, modern, zero-dependency JSON:API client built on the native Fetch API.

fetchja Latest Version Download

Download Latest Version (.zip)

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

# Fetchja

**A tiny, modern, zero-dependency JSON:API client built on the native Fetch API.**

Fetchja helps you talk to a [JSON:API](https://jsonapi.org) server. It uses the browser's own `fetch`, so it stays small and fast. You write plain objects, and Fetchja turns them into JSON:API requests. The server answers, and Fetchja turns the answer back into plain objects — with relationships already filled in. It was inspired by [Kitsu](https://github.com/wopian/kitsu), but it has **no dependencies**.

```js
import Fetchja from 'fetchja'

const api = new Fetchja({
  baseURL: 'https://api.example.com'
})

const { data } = await api.get('articles')
```

## Contents

- [Why Fetchja?](#why-fetchja)
- [Install](#install)
- [Quick start](#quick-start)
- [Options](#options)
- [Methods](#methods)
- [Relationships and included data](#relationships-and-included-data)
- [Meta, links, and the rest of JSON:API](#meta-links-and-the-rest-of-jsonapi)
- [Query parameters](#query-parameters)
- [Make your own query formatter](#make-your-own-query-formatter)
- [Custom fetch](#custom-fetch)
- [Errors](#errors)
- [Retry on error](#retry-on-error)
- [Extensions](#extensions)
- [Atomic Operations](#atomic-operations)
- [Write your own extension](#write-your-own-extension)
- [Use with TanStack Query](#use-with-tanstack-query)
- [TypeScript](#typescript)
- [Plurals](#plurals)

## Why Fetchja?

- ⚡️ **No dependencies.** It only uses the built-in `fetch`. Nothing extra to download. About 6 KB when minified.
- 🧩 **Typed.** It is written in TypeScript and ships its own types. You get autocomplete out of the box.
- 🔄 **Less boilerplate.** You send and read plain objects. Fetchja does the JSON:API parts for you.
- 🪶 **Modern and small.** ESM only, `async`/`await`, no base class to extend.
- 🛡️ **Safe.** It guards against prototype pollution when it reads a response.
- 🎛️ **Flexible.** Set your own headers, query format, name case, plurals, `fetch`, and error handling.

## Install

```bash
npm install fetchja
```

You don't need any other package. Fetchja needs a place where `fetch` exists: Node 18+, Deno, Bun, or any modern browser.

## Quick start

```js
import Fetchja from 'fetchja'

const api = new Fetchja({
  baseURL: 'https://api.example.com'
})

// GET /articles
const { data, meta } = await api.get('articles')

// GET /articles/1
const { data: article } = await api.get('articles/1')

console.log(article)
// { type: 'articles', id: '1', title: 'Hello world', ... }
```

## Options

You set everything when you create the client:

```js
const api = new Fetchja({ /* options */ })
```

| Option | Type | Default | What it does |
| --- | --- | --- | --- |
| `baseURL` | `string` | — | The start of every URL. You need it (or pass one per request). |
| `headers` | `Record<string, string>` | JSON:API headers | Headers added to every request. The JSON:API `Accept` and `Content-Type` are set for you. |
| `fetch` | `typeof fetch` | global `fetch` | Your own fetch function (for example, one that adds a token). |
| `queryFormatter` | `(params) => string \| URLSearchParams` | built-in | Turns the `params` object into a query string. |
| `resourceCase` | `'camel' \| 'kebab' \| 'snake' \| 'none'` | `'none'` | How the resource name in the URL is written. |
| `typeCase` | `'camel' \| 'kebab' \| 'snake' \| 'none'` | `'camel'` | How `type` names are written when you send data. |
| `pluralize` | `boolean \| ((word) => string)` | `true` (built-in) | Make resource names plural. Use `false` to turn it off, or pass your own function. |
| `onResponseError` | `(response) => Response \| void` | — | Runs when a request fails, before Fetchja throws. You can retry here. |

## Methods

| Method | Alias | How to call it | HTTP |
| --- | --- | --- | --- |
| `get` | `fetch` | `get(model, options?)` | `GET` |
| `post` | `create` | `post(model, body, options?)` | `POST` |
| `patch` | `update` | `patch(model, body, options?)` | `PATCH` |
| `delete` | `remove` | `delete(model, id, options?)` | `DELETE` |

Pick the name you like. `api.get` and `api.fetch` do the same thing. So do `create`/`post`, `update`/`patch`, and `remove`/`delete`.

```js
// Read a list
const { data } = await api.get('articles')

// Read one
const { data } = await api.get('articles/1')

// Create
await api.create('article', { title: 'Hello world' })

// Update (put the id in the body)
await api.update('article', { id: '1', title: 'New title' })

// Delete
await api.remove('article', '1') // DELETE /articles/1 (no body)
```

Every call gives you back the data plus a few extra fields:

```js
const response = await api.get('articles/1')

response.data        // your data (an object, or an array for a list)
response.meta        // the JSON:API `meta`, if the server sent it
response.links       // the document links, including pagination
response.jsonapi     // the server's `jsonapi` object
response.status      // 200
response.statusText  // 'OK'
response.headers     // the response headers, as a plain object
```

## Relationships and included data

To send a relationship, put an object (or a list of objects) with a `type` and an `id` inside your data. Fetchja moves it to the right place:

```js
await api.create('article', {
  title: 'Hello world',
  author: { type: 'people', id: '9' },
  tags: [
    { type: 'tags', id: '1' },
    { type: 'tags', id: '2' }
  ]
})
```

An object that is nothing but a `type` and an `id` only links an existing resource, so the request carries no `included` at all. Give it fields and the full resource goes to the top-level `included` too, ready for a server that creates related resources in one request:

```js
await api.create('article', {
  title: 'Hello world',
  author: { type: 'people', id: '9', name: 'Ann' } // also in `included`
})
```

To clear a relationship, send it empty: an empty list for a to-many, and an `id` of `null` for a to-one. Both become `data: null` / `data: []` in the request, the way JSON:API asks for it:

```js
await api.update('article', {
  id: '1',
  author: { id: null }, // relationships.author.data = null
  tags: [] // relationships.tags.data = []
})
```

A plain `author: null` is still an attribute, not a relationship — Fetchja cannot tell the two apart without the object. An object (or a list of objects) with **no** identifier member at all — no `type`, `id`, or `lid` — is a plain JSON attribute and goes under `attributes` as it is, so a JSON column like `metadata: { locale: 'en' }` just works. But an object that carries a `type` and forgets its `id` still throws, so a half-written relationship never clears one by accident. A list of plain values (`tags: ['a', 'b']`) stays an attribute too.

When you read data back, Fetchja takes the resources from `included` and puts them right inside your data. So you can read a relationship like a normal nested object:

```js
const { data } = await api.get('articles/1', {
  params: { include: 'author' }
})

console.log(data.author.name) // comes from `included`
```

## Meta, links, and the rest of JSON:API

A resource can carry more than attributes and relationships: `meta`, `links`, a client-generated `lid`, and the `meta` and `links` of each relationship. None of that fits the flat shape, so Fetchja keeps it all under a single `$` key:

```js
const { data } = await api.get('articles/1')

data.title              // an attribute, flat as always
data.$.meta             // the resource `meta`
data.$.links.self       // the resource `links`
data.$.relationships.author.links.related
data.$.relationships.author.meta
```

Why `$`? JSON:API forbids `$` in member names, so it can never clash with one of your fields. An attribute named `meta` stays exactly where you expect it (`data.meta`), and the resource's own `meta` is still there, next to it (`data.$.meta`). Resources that carry nothing extra have no `$` at all.

Sending works the same way. Anything you put in `$` is written straight into the resource object:

```js
await api.create('article', {
  title: 'Hello world',
  author: { type: 'people', id: '9' },
  $: {
    lid: 'draft-1',
    meta: { draft: true },
    relationships: {
      author: { meta: { role: 'primary' } }
    }
  }
})
```

The `meta` of a relationship's identifiers lives in the raw linkage, so Fetchja keeps that too, in `$.relationships.<name>.data` — the resolved resource in `data.author` stays shared and untouched:

```js
data.tags[0].title                          // resolved from `included`
data.$.relationships.tags.data[0].meta      // the identifier's own `meta`
```

For the document itself, pass `document` with the request:

```js
await api.create('article', { title: 'Hello world' }, {
  document: { meta: { source: 'cli' } }
})
```

## Query parameters

Pass a `params` object. Fetchja serializes it into [JSON:API 1.1](https://jsonapi.org/format/1.1/#query-parameters) query strings: nested objects become bracketed keys, and arrays become comma-separated values (the format the spec defines for `include`, `sort`, `fields[type]`, and list filters).

```js
const { data, meta } = await api.get('articles', {
  params: {
    include: ['author', 'comments'],
    fields: { articles: ['title', 'body'] },
    filter: { published: true },
    sort: ['-createdAt'],
    page: { number: 1, size: 10 }
  }
})
```

This becomes:

```
/articles?include=author,comments&fields[articles]=title,body&filter[published]=true&sort=-createdAt&page[number]=1&page[size]=10
```

Strings work too, so you can pass the comma-separated form directly if you prefer — `include: 'author,comments'` and `sort: '-createdAt'` produce the same output.

Filters of any depth are supported. Object values nest with brackets, scalar arrays join with commas, and arrays of objects (e.g. boolean groups) expand into indexed keys:

```js
await api.get('articles', {
  params: {
    filter: {
      id: [1, 2, 3],            // filter[id]=1,2,3
      price: { gte: 10, lte: 100 }, // filter[price][gte]=10&filter[price][lte]=100
      tags: { any: ['news', 'tech'] }, // filter[tags][any]=news,tech
      or: [{ status: 'active' }, { status: 'pending' }]
      // filter[or][][status]=active&filter[or][][status]=pending
    }
  }
})
```

## Make your own query formatter

The default formatter is JSON:API 1.1 compliant. Some servers want a different shape — for example, repeated `include[]=author&include[]=comments` keys instead of one comma-separated value. You can pass your own function:

```js
import Fetchja from 'fetchja'

function bracketQueryFormatter (params) {
  const search = new URLSearchParams()

for (const key in params) {
    const value = params[key]

if (Array.isArray(value)) {
      for (const item of value) search.append(`${key}[]`, String(item))
    } else {
      search.append(key, String(value))
    }
  }

return search
}

const api = new Fetchja({
  baseURL: 'https://api.example.com',
  queryFormatter: bracketQueryFormatter
})

await api.get('articles', { params: { include: ['author', 'comments'] } })
// -> /articles?include[]=author&include[]=comments
```

Your function takes the `params` object and returns a string or a `URLSearchParams`.

## Custom fetch

You can swap in your own fetch function. This is handy for tokens, timeouts, or logging:

```js
const api = new Fetchja({
  baseURL: 'https://api.example.com',
  fetch: (url, init) => myCustomFetch(url, init)
})
```

## Errors

When a request fails, Fetchja throws a `FetchjaError`. It holds the HTTP status and the JSON:API `errors` list:

```js
import Fetchja, { FetchjaError } from 'fetchja'

try {
  await api.get('articles/999')
} catch (error) {
  if (error instanceof FetchjaError) {
    console.log(error.status)     // 404
    console.log(error.statusText) // 'Not Found'
    console.log(error.errors)     // [{ status: '404', detail: 'Not found' }]
    console.log(error.document)   // the whole error document, with its `meta`
    console.log(error.response)   // the raw Response
  }
}
```

## Retry on error

`onResponseError` runs when the server sends a failing response, before Fetchja throws. The response also gets a `replayRequest()` method. It sends the same request again. This is great for refreshing a token and trying once more:

```js
const api = new Fetchja({
  baseURL: 'https://api.example.com',
  headers: { Authorization: `Bearer ${getToken()}` },
  onResponseError: async (response) => {
    if (response.status === 401) {
      api.headers.Authorization = `Bearer ${await refreshToken()}`

return response.replayRequest() // try again with the new token
    }
  }
})
```

Return a `Response` (like the one from `replayRequest()`) to keep going. Return nothing, and Fetchja throws the `FetchjaError`.

## Extensions

JSON:API has [extensions](https://jsonapi.org/format/#extensions): named add-ons that put extra members in a document. Fetchja ships the official ones as separate imports, so you only pay for the one you use. Register them with the `extensions` option:

```js
import Fetchja from 'fetchja'
import { AtomicOperations } from 'fetchja/atomic'

const api = new Fetchja({
  baseURL: 'https://api.example.com',
  extensions: [AtomicOperations]
})
```

Pass an extension bare, or call it to configure it:

```js
extensions: [AtomicOperations({ endpoint: 'bulk' })]
```

An extension can add methods to the client, rewrite a request before it goes out, and read the response before you see it. Everything else keeps working exactly as it did.

## Atomic Operations

[Atomic Operations](https://jsonapi.org/ext/atomic/) sends an ordered, all-or-nothing batch of writes to a single endpoint. Either every operation lands, or none of them do.

`api.atomic()` takes a callback. It hands you a builder, and you return the operations:

```js
import { AtomicOperations } from 'fetchja/atomic'

const api = new Fetchja({
  baseURL: 'https://api.example.com',
  extensions: [AtomicOperations]
})

const { results } = await api.atomic(op => [
  op.add('author', { lid: 'a1', name: 'dgeb' }),
  op.add('article', {
    title: 'JSON API paints my bikeshed!',
    author: { type: 'authors', lid: 'a1' }
  }),
  op.update('article', { id: '13', title: 'To TDD or Not' }),
  op.remove('article', '9')
])
```

That posts to `/operations` with `Accept` and `Content-Type` set to `application/vnd.api+json;ext="https://jsonapi.org/ext/atomic"`. Only this request carries them, so the rest of your calls stay plain JSON:API.

The builder has three verbs and an escape hatch:

```js
op.add('article', { title: 'Hello' })              // create
op.update('article', { id: '13', title: 'Hi' })    // update, `id` goes in the body
op.remove('article', '13')                         // delete
op.raw({ op: 'add', href: '/anything', data })     // sent untouched
```

Each one takes a last `options` argument of `{ href, meta }`. `href` targets an endpoint of the server's choosing, and it cannot be combined with a `ref`.

### Relationships

Give a `ref` object instead of a model name to target a relationship:

```js
const author = { type: 'articles', id: '13', relationship: 'author' }
const comments = { type: 'articles', id: '1', relationship: 'comments' }

await api.atomic(op => [
  op.update(author, { type: 'people', id: '9' }),           // set a to-one
  op.update(author, null),                                  // clear it
  op.add(comments, [{ type: 'comments', id: '123' }]),      // add members
  op.remove(comments, [{ type: 'comments', id: '12' }])     // drop members
])
```

A `ref` is passed through as written, so its `type` is the one that goes on the wire. A model name (`'article'`) gets the usual casing and pluralization.

### Linking new resources

An operation can name a resource that does not exist yet. Give it a `lid`, then point at that same `lid` from anywhere in the batch:

```js
await api.atomic(op => [
  op.add('author', { lid: 'a1', name: 'dgeb' }),
  op.add('article', { title: 'Hello', author: { type: 'authors', lid: 'a1' } })
])
```

An operation cannot sidepost. Nesting a whole related resource inside another one throws, because an operation has nowhere to put it. Give the related resource its own `add` and link the two by `lid`, as above.

### The response

`results` is positional: one entry per operation, in the order you sent them. An operation the server answers without data gives `null`:

```js
const { results, meta, status } = await api.atomic(op => [
  op.add('author', { name: 'dgeb' }),
  op.remove('author', '8')
])

results // [{ type: 'authors', id: '9', name: 'dgeb' }, null]
```

You also get `document` (the untouched response document), `links`, `jsonapi`, `statusText`, and `headers`. A `204 No Content` gives you `results: []`.

A failed batch is an ordinary `FetchjaError`, and the server points at the operation that broke it:

```js
error.errors[0].source.pointer // '/atomic:operations/1/data/attributes/title'
```

In TypeScript, importing `fetchja/atomic` types `atomic` on every `Fetchja` instance, including ones built without the extension. Calling it there is a plain `TypeError`.

## Write your own extension

An extension is an object with a `name` and whatever else it needs:

```js
const timing = {
  name: 'timing',

methods: (api) => ({
    ping: () => api.get('health')
  }),

onRequest: (context) => {
    context.headers['X-Sent-At'] = String(Date.now())
    context.url.searchParams.set('traced', '1')
  },

onResponse: (payload, response) => {
    // return an object to replace the payload
  },

onError: (error) => {
    console.warn(error.status)
  }
}
```

Hooks run in registration order. `methods` are installed on the instance, and an extension that tries to overwrite a built-in one (`get`, `request`, …) throws instead.

If your extension speaks a media type of its own, `jsonApiMediaType` builds it:

```js
import { jsonApiMediaType } from 'fetchja'

jsonApiMediaType({ ext: ['https://jsonapi.org/ext/atomic'] })
// -> 'application/vnd.api+json;ext="https://jsonapi.org/ext/atomic"'
```

Fetchja never adds those parameters to your everyday requests. Set them on the requests your extension makes, so a server that has never heard of it keeps answering the rest.

## Use with TanStack Query

Fetchja works well with [TanStack Query](https://tanstack.com/query). Because Fetchja throws on a failing request, TanStack Query sees the error and handles it for you.

Reading data:

```js
import { useQuery } from '@tanstack/react-query'
import Fetchja from 'fetchja'

const api = new Fetchja({ baseURL: 'https://api.example.com' })

function useArticles () {
  return useQuery({
    queryKey: ['articles'],
    queryFn: () => api.get('articles')
  })
}
```

Creating data:

```js
import { useMutation, useQueryClient } from '@tanstack/react-query'

function useCreateArticle () {
  const queryClient = useQueryClient()

return useMutation({
    mutationFn: (article) => api.create('article', article),
    onSuccess: () => {
      queryClient.invalidateQueries({ queryKey: ['articles'] })
    }
  })
}
```

## TypeScript

Fetchja is written in TypeScript and brings its own types. You don't need an extra `@types` package.

```ts
import Fetchja, { type FetchjaOptions, FetchjaError } from 'fetchja'

const options: FetchjaOptions = {
  baseURL: 'https://api.example.com',
  resourceCase: 'kebab'
}

const api = new Fetchja(options)
```

The JSON:API types are exported too, so you can type the `$` envelope and everything inside it:

```ts
import type {
  JsonApiDocument,
  JsonApiLinks,
  JsonApiMeta,
  JsonApiRelationship,
  ResourceEnvelope,
  ResourceIdentifier
} from 'fetchja'

interface Article {
  type: 'articles'
  id: string
  title: string
  $?: ResourceEnvelope
}
```

## Plurals

By default, Fetchja makes resource names plural with a small built-in helper. It knows the common English rules (and `status` becomes `statuses`, not `statu`). It is also safe to run twice: `articles` stays `articles`.

For tricky words (like `person` → `people`), pass a bigger library such as [`pluralize`](https://www.npmjs.com/package/pluralize):

```js
import pluralize from 'pluralize'

const api = new Fetchja({
  baseURL: 'https://api.example.com',
  pluralize
})
```

Or turn it off and use the exact names you write:

```js
const api = new Fetchja({
  baseURL: 'https://api.example.com',
  pluralize: false
})
```

fetchja

fetchja Latest Version Download

Today's picks

Was this content helpful?

Recent Feedback

fetchja

fetchja Latest Version Download

Today's picks

Was this content helpful?

Recent Feedback

Download README