> ## 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.

# Read text aloud

> Read Japanese text aloud with Edge TTS or Gemini TTS, with adjustable voice and speed.

You can have the app read the Japanese text in the input box aloud, or play the pronunciation of a single word in the word details panel.

## Read the text aloud

Click the **Listen** button in the input area. When you hover over it, the button shows the estimated time, for example “Read aloud (about 5–10 seconds)”.

| Text length              | Estimated time |
| ------------------------ | -------------- |
| Up to 20 characters      | 5–10 seconds   |
| 21–50 characters         | 10–20 seconds  |
| 51–100 characters        | 20–30 seconds  |
| More than 100 characters | 30–60 seconds  |

In the word details panel, click **Play pronunciation** to read only the selected word.

## Choose a speech engine

Open **Voice settings** in the input area and choose an option under **Speech engine**:

<Tabs>
  <Tab title="Edge TTS (default)">
    Works out of the box with no API key.

    | Setting | Options                                                                                                         |
    | ------- | --------------------------------------------------------------------------------------------------------------- |
    | Voice   | **Female (Nanami)**: `ja-JP-NanamiNeural`<br />**Male**: `ja-JP-KeitaNeural`                                    |
    | Speed   | From -100 to +100 in steps of 10. The interface shows the value as Very slow, Slow, Normal, Fast, or Very fast. |
  </Tab>

  <Tab title="Gemini TTS">
    Uses the `gemini-3.1-flash-tts-preview` model and requires a Gemini API key. It works if the server has `GEMINI_API_KEY` configured, or if you entered a Gemini key in **Settings**.

    | Voice  | Style       |
    | ------ | ----------- |
    | Kore   | Firm        |
    | Puck   | Upbeat      |
    | Zephyr | Bright      |
    | Aoede  | Breezy      |
    | Leda   | Youthful    |
    | Charon | Informative |

    **Speaking style** options:

    * **Natural**: The default style.
    * **Slow**: Slows down the speech, which helps with shadowing practice.
    * **Clear**: Pronounces words more clearly.
  </Tab>
</Tabs>

Your speech engine, voice, speed, and style are saved in the current browser and restored automatically the next time you open the app.

## When text-to-speech fails

| Message                                                     | Fix                                                                                           |
| ----------------------------------------------------------- | --------------------------------------------------------------------------------------------- |
| Enter some text first                                       | The input box is empty. Enter the Japanese you want to hear first.                            |
| No API key provided                                         | You're using Gemini TTS without a usable Gemini key. Enter a key, or switch back to Edge TTS. |
| The upstream TTS request timed out. Please try again later. | The speech service didn't respond within 60 seconds. Shorten the text and try again.          |
| Edge TTS request failed (HTTP status code)                  | The Edge TTS speech API returned an error. Try again later, or switch to Gemini TTS.          |
| Edge TTS returned empty audio                               | The speech service returned no audio. Try again later, or switch to Gemini TTS.               |

<Note>
  Edge TTS requests are forwarded through the app server to the speech API `api.howen.ink`, provided by the project author. If you don't want your text sent to that service, use Gemini TTS instead. For details, see [Keys and privacy](/en/privacy).
</Note>