> ## Documentation Index
> Fetch the complete documentation index at: https://doc.howen.ink/llms.txt
> Use this file to discover all available pages before exploring further.

# 日本語文章解析

> Read Japanese sentences word by word. Type or paste Japanese to see word segmentation, furigana, parts of speech, meanings, and a full translation.

Japanese Text Analyzer (日本語文章解析) is an AI-powered web app for breaking down Japanese sentences. You type or paste Japanese text, and it labels each word with its kana reading, romaji, and part of speech. A full translation appears below. Click any word to see its meaning in that sentence, its usage, its conjugation, and an example sentence.

<Info>
  Want to try it first? Open the [live demo](https://nihongodemo.howen.ink/) and start analyzing without deploying anything.
</Info>

![Analysis results with dictionary definition](https://raw.githubusercontent.com/cokice/japanese-analyzer/master/docs/images/app-home.png)

## Features at a glance

### Read sentences

<CardGroup cols={3}>
  <Card title="Sentence analysis" icon="spell-check" href="/en/features/analysis">
    Label each word with furigana, romaji, and part of speech, and get a full translation automatically.
  </Card>

  <Card title="Word details" icon="book-open" href="/en/features/word-detail">
    Click a word to see its meaning in the sentence, dictionary form, conjugation, and an example.
  </Card>

  <Card title="Multi-word selection" icon="highlighter" href="/en/features/word-detail#multi-word-selection">
    Get grammar patterns and set expressions explained as a whole. Merge a wrongly split word in one click.
  </Card>
</CardGroup>

### Work more smoothly

<CardGroup cols={2}>
  <Card title="Sentence of the day and recent texts" icon="calendar-days" href="/en/features/daily-sentence">
    The home page shows a new sentence each day, updated on Japan time. Your recent analyses are saved locally in your browser.
  </Card>

  <Card title="Image recognition" icon="image" href="/en/features/image-recognition">
    Upload or paste a screenshot to extract the Japanese text in it.
  </Card>

  <Card title="Read text aloud" icon="volume-high" href="/en/features/text-to-speech">
    Listen with Edge TTS or Gemini TTS, with adjustable speed and voice.
  </Card>

  <Card title="AI Japanese Assistant" icon="comments" href="/en/features/ai-assistant">
    Ask follow-up questions about grammar, vocabulary, and learning, in the context of the current sentence.
  </Card>
</CardGroup>

### Interface

* Supports 简体中文, 繁體中文, English, and 한국어. The interface, translations, and definitions switch together.
* Offers three themes: light, dark, and system. Works on desktop and mobile.
* Splits long texts into chunks automatically. When you paste from a web page or Markdown, formatting and links are removed automatically.

<Columns cols={2}>
  <Frame caption="Dark mode">
    <img src="https://raw.githubusercontent.com/cokice/japanese-analyzer/master/docs/images/app-dark.png" alt="Analysis interface in dark mode" />
  </Frame>

  <Frame caption="AI Japanese Assistant on a phone">
    <img src="https://raw.githubusercontent.com/cokice/japanese-analyzer/master/docs/images/mobile-chat.png" alt="AI Japanese Assistant on a phone" />
  </Frame>
</Columns>

## Models

| Purpose                                          | Default                         | Options                                                   |
| ------------------------------------------------ | ------------------------------- | --------------------------------------------------------- |
| Analysis, translation, definitions, AI assistant | DeepSeek `deepseek-flash`       | Gemini `gemini-flash-latest` / `gemini-flash-lite-latest` |
| Image recognition                                | Same as the selected text model | —                                                         |
| Text-to-speech                                   | Edge TTS                        | Gemini TTS (requires a Gemini API key)                    |

The API key configured on the server is shared by all visitors. Users can also enter their own key in **Settings**. That key stays in their own browser, and the app server forwards it to the model provider with each request. For details, see [Keys and privacy](/en/privacy).

## Tech stack

* Next.js 15 (App Router) and React 19
* Tailwind CSS
* DeepSeek and Gemini through OpenAI-compatible APIs
* License: [AGPL-3.0-only](/en/license)

## Next steps

<Columns cols={2}>
  <Card title="Quickstart" icon="rocket" href="/en/quickstart">
    Install dependencies locally and run the app in minutes.
  </Card>

  <Card title="Deployment" icon="server" href="/en/deployment">
    Deploy to Vercel, or run it on your own server with Docker.
  </Card>

  <Card title="Deploy with an AI agent" icon="robot" href="/en/ai-deploy">
    Hand deployment to an AI coding assistant such as Claude Code or Codex with a single sentence.
  </Card>

  <Card title="Configuration" icon="sliders" href="/en/configuration">
    Learn about every environment variable, the models, and the access password.
  </Card>
</Columns>