Getting Started

Installation

With fetch (built into core — zero extra dependencies):

npm install @routar/core

With Axios:

npm install @routar/core @routar/axios axios

Set up with an AI agent

Paste this into Cursor, Claude Code, or any AI coding assistant:

Add routar (@routar/core) as the API client for this project. First fetch the latest routar docs — via Context7 (use context7) if that MCP server is set up, otherwise fetch https://routar.vercel.app/llms-full.txt directly. Check package.json for an existing HTTP client: use @routar/axios if axios is already a dependency, @routar/ky if ky is, otherwise use the built-in fetch executor. Then set it up: define Zod schemas for the API models, group endpoints with endpoint() + defineRouter(), and build the typed client with the chosen executor + createApi(). If @tanstack/react-query is already a dependency, don’t wire up @routar/react-query yet — ask me first, then follow up as a separate step, since it touches query keys and provider setup across the app.

Quick Start

1. Define your schema


import { z } from 'zod'
 
const TodoSchema = z.object({
  id: z.number(),
  title: z.string(),
  completed: z.boolean(),
})

2. Define endpoints and router


import { endpoint, defineRouter } from '@routar/core'
 
const todoRouter = defineRouter('/todos', {
  getList: endpoint({
    method: 'GET',
    path: '/',
    response: z.array(TodoSchema),
  }),
  getDetail: endpoint({
    method: 'GET',
    path: '/:id',
    request: { path: z.object({ id: z.number() }) },
    response: TodoSchema,
  }),
  create: endpoint({
    method: 'POST',
    path: '/',
    request: { body: z.object({ title: z.string() }) },
    response: TodoSchema,
  }),
})

3. Create executor and API client


import { createApi } from '@routar/core'
import { createFetchExecutor } from '@routar/core'
 
const executor = createFetchExecutor('https://api.example.com')
const todoApi  = createApi(executor, todoRouter)

4. Make requests


const todos = await todoApi.getList({})
const todo  = await todoApi.getDetail({ path: { id: 1 } })
const next  = await todoApi.create({ body: { title: 'buy milk' } })

All return types are fully inferred — no type annotations needed. In your IDE, hovering over todos shows Todo[] and todo shows Todo.

AbortSignal

Pass an AbortSignal as the second argument to cancel in-flight requests:


const controller = new AbortController()
const todos = await todoApi.getList({}, controller.signal)
 
controller.abort() // cancels the request

With Plugins


import { createFetchExecutor, logger } from '@routar/core'
 
const executor = createFetchExecutor('https://api.example.com', {
  plugins: [logger()],
  retry: 3,
  timeout: 8_000,
})

With Axios


import axios from 'axios'
import { createAxiosExecutor } from '@routar/axios'
 
const executor = createAxiosExecutor(
  axios.create({ baseURL: 'https://api.example.com' })
)

Next Steps

Example app — a full Next.js app using every package, SSR + CSR, and TanStack Query
API Reference — full documentation for every export
Executors — fetch vs. axios, SSR/CSR patterns
Guides — SSR/CSR, custom executors, error handling