End-to-End Type Safety

One of Kanjou's standout features is its End-to-End (E2E) Type Safety. In most i18n libraries, translation keys are weakly typed, and variable arguments are often just Record<string, any>, leading to runtime crashes when variables are missing or misnamed.

Kanjou eliminates this entire class of bugs.

How It Works

1. Locale Analysis: You write your translations in your locale files using MessageFormat 2.
2. Code Generation: You run the Kanjou CLI (@kanjou/cli). The CLI parses your default locale file, extracts all keys, uncovers the expected variables, and identifies formatting functions (like :number).
3. Module Augmentation: The CLI generates a TypeScript definition file (kanjou.d.ts) that augments the global @kanjou/react module.
4. Strict Usage: Your React components now get full IDE autocompletion for keys and variables.

The Developer Experience

Before

// ❌ No autocomplete for "greeting.hello"
// ❌ No warning if you misspell "name" or forget it entirely
const message = t('greeting.hello', { nme: 'Alice' })

After

import { useKanjou } from '@kanjou/react'

function Welcome() {
  const kanjou = useKanjou()

  // ✅ Full autocomplete for 'greeting.hello'
  // ✅ Type error if `name` is omitted or misnamed
  // ✅ Knows that `count` should be a number based on MF2 `:number` formatting!
  return <h1>{kanjou.t('greeting.hello', { name: 'Alice', count: 5 })}</h1>
}

Because Kanjou knows your schema at build-time, your translations and your code are always perfectly in sync.