vocab-chat

name	vocab-chat
description	Create a vocabulary learning chat MulmoScript with messenger-style animated UI (voiceover approach). Use when the user wants to create vocabulary learning content.
argument-hint	<word> (e.g., layover)
allowed-tools	Read, Write, Edit, Bash, Grep, Glob
user-invocable	true

/vocab-chat — Vocabulary Learning Chat Script

Create a messenger-style vocabulary chat video using MulmoCast's voiceover approach. The video shows an animated chat UI where a Tutor and Student discuss a target vocabulary word.

Approach: 3-section structure — intro beat + animated html_tailwind chat beat with voice_over beats + closing beat.

IMPORTANT: The voice_over group must NOT be the first or last beat. Always add intro/closing beats to avoid audio/video duration mismatch caused by add_bgm_agent's introPadding/outroPadding.

---

Phase 1: Content Generation

Input

• Target word: $ARGUMENTS (e.g., "layover")
• If no word is given, ask the user

Generate conversation

Create a 7-message dialogue between Tutor and Student:

1. Tutor asks a question using the word naturally
2. Student asks what the word means
3. Tutor explains the meaning
4. Student uses the word in a sentence
5. Tutor elaborates or gives context
6. Student responds naturally
7. Tutor encourages further use

Guidelines:

• Natural, conversational English
• Target word highlighted where it naturally appears
• 1-2 sentences per message

---

Phase 2: Create Initial Script (without timing)

Output directory

SCRIPTS_DIR="${MULMO_SCRIPTS_DIR:-my-scripts}"

Write the script to ${SCRIPTS_DIR}/test_vocab_chat_voiceover_{word}.json.

Template structure

{
  "$mulmocast": { "version": "1.1" },
  "lang": "en",
  "title": "Vocabulary Chat: {word} (voiceover)",
  "canvasSize": { "width": 720, "height": 1280 },
  "speechParams": {
    "speakers": {
      "Tutor": { "voiceId": "shimmer", "displayName": { "en": "Tutor" } },
      "Student": { "voiceId": "echo", "displayName"     
    
  
   
    
       
       
       
       
         
         
      
    
    
       
       
       
       
         
         
         
         
      
    
    
       
       
       
          
    
    
    
       
       
       
       
         
         
      
    
  

IMPORTANT: In the initial script:

• Do NOT set duration on the intro, chat, or closing beats
• Do NOT set startAt on voice_over beats
• Set showAt to placeholder zeros: var showAt = [0, 0, 0, 0, 0, 0, 0];

Intro beat template

<div class='h-full flex flex-col items-center justify-center bg-gradient-to-br from-slate-900 to-indigo-950'>
  <p class='text-cyan-300 text-2xl mb-4'>Today's Word</p>
  <h1 class='text-5xl font-bold text-white tracking-wide'>{word}</h1>
  <div class='h-1 w-48 bg-gradient-to-r from-cyan-400 to-indigo-400 mt-6 rounded'></div>
</div>

Closing beat template

<div class='h-full flex flex-col items-center justify-center bg-gradient-to-br from-slate-900 to-indigo-950'>
  <p class='text-3xl text-white font-bold mb-4'>Great job!</p>
  <p class='text-2xl text-cyan-300'>You learned: <span class='font-semibold'>{word}</span></p>
  <p class='text-xl text-slate-400 mt-6'>Keep practicing in your daily conversations.</p>
</div>

Chat HTML Template — English Only

<div class='h-full flex flex-col bg-slate-900'>
  <div class='bg-gradient-to-r from-cyan-600 to-indigo-600 px-6 py-4'>
    <p class='text-white text-3xl font-bold'>Vocabulary Chat</p>
    <p class='text-cyan-100 text-2xl'>Learning: {word}</p>
  </div>
  <div class='flex-1 px-4 py-4 flex flex-col gap-3'>
    <!-- m1 through m7 message bubbles -->
  </div>
</div>

Chat HTML Template — With Translation

When translation is enabled, add a t{N} element inside each message:

Tutor message (left-aligned, with translation):

<div id='m{N}' class='flex gap-2 items-end' style='opacity:0'>
  <div class='w-7 h-7 bg-cyan-500 rounded-full flex items-center justify-center text-white text-xs font-bold shrink-0'>T</div>
  <div>
    <div class='bg-slate-700 rounded-2xl rounded-bl-sm px-4 py-2.5 max-w-[75%]'>
      <p class='text-white text-2xl leading-relaxed'>English with <span class='text-cyan-400 font-semibold'>{word}</span></p>
    </div>
    <p id='t{N}' class='text-slate-500 text-2xl mt-0.5 ml-2' style='opacity:0'>日本語訳</p>
  </div>
</div>

Student message (right-aligned, with translation):

<div id='m{N}' class='flex gap-2 items-end justify-end' style='opacity:0'>
  <div>
    <div class='bg-indigo-600 rounded-2xl rounded-br-sm px-4 py-2.5 max-w-[75%]'>
      <p class='text-white text-2xl leading-relaxed'>English with <span class='text-indigo-300 font-semibold'>{word}</span></p>
    </div>
    <p id='t{N}' class='text-slate-500 text-2xl mt-0.5 mr-2 text-right' style='opacity:0'>日本語訳</p>
  </div>
  <div class='w-7 h-7 bg-indigo-400 rounded-full flex items-center justify-center text-white text-xs font-bold shrink-0'>S</div>
</div>

Animation script (image.script field)

The animation code goes in a separate "script" field (not inside html).

For English only:

"script": [
  "function render(frame, totalFrames, fps) {",
  "  var showAt = [0, 0, 0, 0, 0, 0, 0];",
  "  for (var i = 0; i < 7; i++) {",
  "    var el = document.getElementById('m' + (i + 1));",
  "    var start = showAt[i];",
  "    el.style.opacity = interpolate(frame, {input:{inMin:start,inMax:start+12},output:{outMin:0,outMax:1},easing:'easeOut'});",
  "    el.style.transform = 'translateY(' + interpolate(frame, {input:{inMin:start,inMax:start+12},output:{outMin:20,outMax:0},easing:'easeOut'}) + 'px)';",
  "  }",
  "}"
]

For English + Translation:

"script": [
  "function render(frame, totalFrames, fps) {",
  "  var showAt = [0, 0, 0, 0, 0, 0, 0];",
  "  var showAtJa = [0, 0, 0, 0, 0, 0, 0];",
  "  for (var i = 0; i < 7; i++) {",
  "    var el = document.getElementById('m' + (i + 1));",
  "    var start = showAt[i];",
  "    el.style.opacity = interpolate(frame, {input:{inMin:start,inMax:start+12},output:{outMin:0,outMax:1},easing:'easeOut'});",
  "    el.style.transform = 'translateY(' + interpolate(frame, {input:{inMin:start,inMax:start+12},output:{outMin:20,outMax:0},easing:'easeOut'}) + 'px)';",
  "    var tl = document.getElementById('t' + (i + 1));",
  "    if (tl) {",
  "      var tstart = showAtJa[i];",
  "      tl.style.opacity = interpolate(frame, {input:{inMin:tstart,inMax:tstart+12},output:{outMin:0,outMax:1},easing:'easeOut'});",
  "    }",
  "  }",
  "}"
]

---

Phase 3: Generate Audio & Calculate Timing

Step 1: Generate audio

yarn audio ${SCRIPTS_DIR}/test_vocab_chat_voiceover_{word}.json

Step 2: Run timing calculator

For English only:

npx tsx .claude/skills/vocab-chat/calc_voiceover_timing.ts ${SCRIPTS_DIR}/test_vocab_chat_voiceover_{word}.json

For English + Translation (1 second delay):

npx tsx .claude/skills/vocab-chat/calc_voiceover_timing.ts ${SCRIPTS_DIR}/test_vocab_chat_voiceover_{word}.json --translation-delay 1.0

The script automatically:

• Reads audio durations from the studio JSON (via ffprobe)
• Calculates startAt for each voice_over beat
• Calculates showAt (and showAtJa) frames for animation
• Sets duration on the parent beat
• Updates the script JSON in-place

Use --dry-run to preview calculations without modifying the file.

---

Phase 4: Generate Movie

yarn movie ${SCRIPTS_DIR}/test_vocab_chat_voiceover_{word}.json

CRITICAL: Never use -f flag after timing is set. TTS generates different audio each time, invalidating the hardcoded timing.

Output: output/test_vocab_chat_voiceover_{word}_en.mp4

---

Beat Structure Overview

Beat	Type	Role
intro	html_tailwind (static)	Word introduction card
chat	html_tailwind (animated) + voice_over parent	Chat UI with 7 messages
vo2–vo7	voice_over	Audio for messages 2–7
closing	html_tailwind (static)	Closing/summary card

Why intro/closing beats are required: MulmoCast's add_bgm_agent adds introPadding (1s) before the first beat's audio and outroPadding (1s) after the last beat's audio. movie.ts adds matching visual padding only to the first and last beats via getExtraPadding. If the voice_over group is at the first/last position, voice_over beats are skipped in video processing, causing the visual padding to be lost — resulting in audio/video duration mismatch. By placing intro/closing beats at the edges, the padding is correctly applied to those static beats.

---

Timing Calculation Details (for reference)

The calc_voiceover_timing.ts script uses these formulas:

Value	Formula
roundUp1(x)	Math.ceil(x * 10) / 10
startAt[i]	roundUp1(startAt[i-1] + audioDuration[i-1] + 0.3)
showAt[0]	Math.round(firstShowAtDelay * 30)
showAt[i] (i>0)	Math.round(startAt[i] * 30)
showAtJa[i]	showAt[i] + Math.round(delay * 30)
duration	roundUp1(startAt[last] + audioDuration[last] + 0.3)

Constants: firstShowAtDelay = 1.0s (first beat) or 0s (middle beat), fps = 30, gap = 0.3s

---

Reference

Timing calculation details: see .claude/skills/vocab-chat/calc_voiceover_timing.ts and the timing formulas above.