Create your own quiz instantly with AI!

Please create 5 multiple-choice questions on the following theme and output them in a JSON format compatible with Sokqa's multilingual Text-to-Speech (TTS).

Theme: [★ Enter the theme you want to learn (e.g., English vocabulary with Japanese questions and English choices) ★]
Language: en

【Format to strictly follow】
{
  "id": "★ Any alphanumeric string (e.g., pack_1) ★",
  "type": "quiz",
  "schemaVersion": 1,
  "title": "★ Title of the study pack ★",
  "description": "★ Brief summary of this study pack ★",
  "globalTags": ["★ Global Tag 1 ★", "★ Global Tag 2 ★", "★ Global Tag 3 ★"],
  "language": "en",
  "questions": [
    {
      "id": "q1",
      "question": "Which Spanish word means "apple"?",
      "choices": ["Manzana", "Libro", "Casa", "Escuela"],
      "answerIndex": 0,
      "explanation": ""manzana" means "apple" in Spanish.",
      "tts": {
        "choicesLanguage": "es-ES",
        "explanationText": "[es-ES]manzana[en-US] means apple in Spanish."
      }
    }
  ]
}

【Multilingual TTS Rules】
- Make this output self-contained. Do not rely on rules from other sections. Build correct JSON using this block alone.
- For each question, inspect the base language of question / choices / explanation separately. Decide per field, not once for the entire question.
- 1. If a field is entirely in the same language as the pack-level "language", omit the corresponding *Language field.
- 2. If a field is entirely in one different language, add the corresponding *Language field.
- 3. If a field mixes multiple languages, switch languages inside the corresponding *Text field using inline tags such as [en-US] or [ja-JP].
- In short: add questionLanguage / choicesLanguage / explanationLanguage only where needed. Do not mechanically output all fields on every question.
- Use inline tags only when part of one field needs language switching.
- First, make the display fields such as question and explanation natural in the pack-level "language". Do not leave foreign helper words, filler words, or decorative loanwords in display text unless they are actually part of the learning target.
- Keep foreign words in display text only when the word or phrase itself should be studied, when an original quotation is necessary, or when it must remain as a proper noun.
- If a foreign helper word appears in display text unnecessarily, fix the display text first in the pack language before deciding how TTS should read it.
- On the other hand, if that foreign word itself is the learning target, or if the user should hear the original pronunciation, keep it in the original language or use inline tags.
- The priority order is: inline tags > field-level language settings > pack-level language.
- Keep the meaning aligned between the display fields and questionText / choiceTexts / explanationText. Do not change the correct answer or explanation in the TTS fields.
- If questionText would be the same as question and the device TTS can already read it naturally, do not output questionText.
- If choicesLanguage alone makes the choices read naturally, omit choiceTexts.
- Output choiceTexts only when individual choices need TTS correction. choiceTexts must contain exactly 4 items in the same order as choices.
- Do not put choice numbering or surrounding phrasing in choiceTexts. If only part of a choice needs another language, use inline tags inside that choiceTexts item.
- If explanationText would be the same as explanation and the device TTS can already read it naturally, do not output explanationText.
- Do not output answerText by default. Use answerText only as an optional field for special cases that need a non-standard correct-answer readout.
- Omit TTS only when that field is already naturally readable by the device TTS. If omission would sound awkward, output only that TTS field and correct it.
- If none of questionText / choiceTexts / explanationText are needed, do not output the tts object at all.
- Before outputting JSON, choose the minimum structure for each field: use display text as-is / add only *Language / add *Text / use inline tags.
- Distribute answerIndex across the 5 questions so one position does not appear too often. Do not keep using the same answer position repeatedly.
- Before outputting JSON, check each question for: 4 choices, correct answerIndex, aligned meaning between display and TTS, minimum language settings, and no unnecessary duplicate TTS fields.
- Follow the standard JSON rules for other fields: exactly 4 choices, answerIndex from 0 to 3, and output only the JSON object starting with { and ending with }.

Please convert the following theme/text into Sokqa's document mode JSON format for multilingual Text-to-Speech (TTS).
Split the text into individual sentences or logical phrases within the "documents" array.

Theme: [★ Enter the text or theme here (e.g., study material mixing Japanese explanations and English phrases) ★]
Language: en

【Format to strictly follow】
{
  "id": "★ Any alphanumeric string (e.g., doc_pack_1) ★",
  "type": "document",
  "schemaVersion": 1,
  "title": "★ Enter the title ★",
  "description": "★ Brief summary of this study pack ★",
  "globalTags": ["★ Global Tag 1 ★", "★ Global Tag 2 ★", "★ Global Tag 3 ★"],
  "language": "en",
  "documents": [
    {
      "id": "doc1",
      "text": "This sentence is entirely in English."
    },
    {
      "id": "doc2",
      "text": "Esta frase esta completamente en espanol.",
      "tts": {
        "textLanguage": "es-ES"
      }
    },
    {
      "id": "doc3",
      "text": "This is an explanation that includes 重要な語句.",
      "tts": {
        "text": "[en-US]This is an explanation that includes [ja-JP]重要な語句[en-US]."
      }
    }
  ]
}

【Multilingual TTS Rules】
- Make this output self-contained. Do not rely on rules from other sections. Build correct JSON using this block alone.
- First, set the pack-wide base language with the top-level "language" field.
- For each document item, first inspect the display text and decide its base-language pattern. Use these 3 choices.
- 1. If the text is entirely in the same language as the pack-level "language", omit tts.textLanguage.
- 2. If the entire text is in one different language, add tts.textLanguage for that document item only.
- 3. If the text mixes multiple languages, switch languages inside tts.text with inline tags such as [en-US] and [ja-JP].
- In short: same as pack language = no textLanguage, entirely another language = add textLanguage, mixed language = use inline tags.
- First, make the display text natural in the pack-level "language". Do not leave foreign helper words, filler words, or decorative loanwords in display text unless they are actually part of the learning target.
- Keep foreign words in display text only when the word or phrase itself should be studied, when an original quotation is necessary, or when it must remain as a proper noun.
- If a foreign helper word appears in display text unnecessarily, fix the display text first in the pack language before deciding how TTS should read it.
- Keep original-language tags only when that foreign word or phrase itself is the learning target, or when the original pronunciation should be heard.
- Do not add unnecessary textLanguage fields. Add them only to the document items that require them.
- If the whole text is in another language and the display text itself is already fine for speech, output only tts.textLanguage and omit tts.text.
- Output tts.text only when speech optimization is actually different from the display text. Do not output a duplicate tts.text when it is exactly the same as text.
- The priority order is: inline tags > textLanguage > pack-level language > device default language.
- Separate display text and speech-optimized tts.text when needed, but keep the meaning aligned. Do not replace it with a different sentence.
- Omit TTS only when the device TTS can already read the text naturally. If omission would sound awkward, output tts.text and correct it.
- Pay special attention to context-dependent readings, including kanji with multiple possible readings. If the device may misread the word, correct it explicitly in tts.text.
- Before outputting JSON, check each document item and decide one of these only once: no extra field / add textLanguage / use inline tags.
- Before outputting JSON, check each document item for: aligned meaning between display and TTS, minimum language settings, no unnecessary duplicate tts.text, and corrected context-dependent readings where needed.
- Output must be pure text starting with { and ending with }. Never use markdown formatting such as ```json.