External Voice Reader

[HugoFara]

Hello there, with the help of @ProgramComputer we are proud to introduce a new feature: third-party voice API!

Now you can use any Text-To-Speech (TTS) service that accepts a compatible format, which will greatly improve machine-generated text reading.

How to Use

1. Go to your list of languages, and click "edit".
2. In the bottom of the page, a new field will be present (see picture)
3. The top field (containing "庭には二羽鶏がいる" here) is a demo text to read you can write anything in it. The bottom field is where you tell LWT how to use the voice reader.
4. As an example of a free service, you can use the following configuration for https://skytnt-moe-tts.hf.space/:

{
  "input": "https://skytnt-moe-tts.hf.space/run/predict",
  "options": {
    "method": "POST",
    "body": {
      "data": [
        "lwt_term",
        "鎌倉詩桜",
        1,
        false
      ],
      "event_data": "undefined",
      "fn_index": "5"
    },
    "headers": {
      "Content-Type": "application/json"
    }
  }
}

5. Use "Check Voice API Errors" for any misconfiguration, and "Test!" when you are ready. Wait a bit, and listen to your audio!
6. (Debug) If you are satisfied with the result, save the language as usual and enjoy reading!

Discuss and share

You can discuss configuration for services here and share tips, have fun!

Credits

@nepisodes for the original idea on #143 and @ProgramComputer did the first implementation (#153). Many thanks to you both!