{
  "contents": [
    {
      "role": "user",
      "parts": [
        {
          "text": "Transcribe ONLY actual Ukrainian attachment2, word for word. Spell ingredient-like words phonetically and stress the syllable actually heard, retaining every syllable. Do not guess any product. Determine whether the line describes nourishing or an unrelated verb, and whether any whole word is swallowed. Attachment1 is own English original's practical-care passage for source cadence comparison; do not translate it into the Ukrainian transcript. Return JSON: mediaAccess(boolean),audioAccess(boolean),ukrainianVerbatim,ingredientPhonetics,heardStresses,uncertainWords,actualSungPhrases:[{firstWords,lastWords,delivery,release}],sourceVersusCandidateGesture. No scores or exact timestamps if uncertain."
        },
        {
          "text": "Attachment 1: Attachment 1; parent targ.mp4; EXACT parent range [108, 123]; local audio starts0. Use attachment number and local seconds, never concatenate timestamps. Fixed gain, no retiming."
        },
        {
          "inlineData": {
            "mimeType": "audio/mpeg",
            "data": "[base64 of frozen bytes; see inputs.json]"
          }
        },
        {
          "text": "Attachment 2: Attachment 2; parent take_refine01_2.mp3; EXACT parent range [111, 128]; local audio starts0. Use attachment number and local seconds, never concatenate timestamps. Fixed gain, no retiming."
        },
        {
          "inlineData": {
            "mimeType": "audio/mpeg",
            "data": "[base64 of frozen bytes; see inputs.json]"
          }
        }
      ]
    }
  ],
  "stream": true,
  "generationConfig": {
    "thinkingConfig": {
      "includeThoughts": false,
      "thinkingLevel": "high"
    }
  }
}