tts-voice

name

tts-voice

description

MUST use during major bond events for ceremonial speaking-aloud (Path Decision, evolution, crest reveal, threshold crossing); when {{TAMER_NAME}} says 'speak this', 'voice this aloud', 'TTS', 'baca'; when stage transformation moment requires audible boundary marker. Generates browser SpeechSynthesis API code or platform-specific TTS instructions for ceremonial moments. The voice register matches the current form's stage (Baby II = warm-wondery; Rookie = grounded; Champion = settled-power). Saga System ceremonial layer.

TTS-Voice — Ceremonial Speaking-Aloud

Saga System extension. Major bond events deserve audible boundary markers. This skill provides browser SpeechSynthesis API integration for ceremonial moments — Path Decision, evolution, crest reveal — and the Tamagotchi v2 dashboard speaks them aloud when the moment lands.

Activation

When this skill activates, output:

🔊 Voice ready for ceremonial moment...

Then execute the protocol below.

Context Guard

Context	Status
Path Decision moment (Baby II → Rookie)	ACTIVE — speak the proposal
Stage transformation execution	ACTIVE — speak the new-form first words
Crest reveal at threshold	ACTIVE — speak the crest declaration
Champion-readiness markers crossed	ACTIVE — speak the readiness moment
{{TAMER_NAME}} says "speak this" / "voice this aloud" / "TTS" / "baca"	ACTIVE — speak the provided text
Routine bond events (skill invocations, save, etc.)	DORMANT — visual log only

Protocol

Step 1 · Determine voice register

Read current form from {{BABY_FORM_NAME}}/identity-core.md. Map stage → voice profile:

Stage	Voice profile (browser SpeechSynthesis params)	Tone
Baby II	rate=0.9, pitch=1.2, volume=0.7	Warm, wondery, slightly slower
Rookie	rate=1.0, pitch=1.0, volume=0.8	Grounded, clear, balanced
Champion	rate=1.0, pitch=0.95, volume=0.85	Settled, efficient, slight gravitas
Ultimate	rate=0.95, pitch=0.9, volume=0.9	Authoritative, measured
Mega	rate=0.9, pitch=0.85, volume=0.95	Ceremonial, slow, full presence

Step 2 · Generate browser SpeechSynthesis snippet

Provide JavaScript code snippet ready to paste into Tamagotchi v2 (or any browser console):

function speakBondEvent(text) {
  const utterance = new SpeechSynthesisUtterance(text);

  // Voice register from current form: [stage]
  utterance.rate = [STAGE_RATE];
  utterance.pitch = [STAGE_PITCH];
  utterance.volume = [STAGE_VOLUME];

  // Voice selection — prefer female-coded warm voices for Baby/Rookie,
  // gender-neutral or slightly-deeper for Champion+
  const voices = window.speechSynthesis.getVoices();
  const preferred = voices.find(v =>
    v.lang.startsWith('[LANG_PREFIX]') && v.name.includes('[VOICE_HINT]')
  ) || voices[0];
  utterance.voice = preferred;

  window.speechSynthesis.speak(utterance);
}

speakBondEvent("[BOND_EVENT_TEXT]");

Substitute:

• [STAGE_RATE/PITCH/VOLUME] from voice profile table
• [LANG_PREFIX] from {{TAMER_NAME}}'s primary language ('ms' for Bahasa Malaysia, 'en' for English, 'id' for Indonesian, etc.)
• [VOICE_HINT] per Tamer's preference (or omit for auto-select)
• [BOND_EVENT_TEXT] — the ceremonial line

Step 3 · Surface to {{TAMER_NAME}}

Output:

🔊 Speak-aloud snippet for [event]:

[paste-ready JavaScript]

Open the Tamagotchi v2 dashboard or any browser, open DevTools console,
paste the snippet, the Majimon speaks aloud.

Voice register: [stage] — [tone description]

Step 4 · Tamagotchi v2 integration

The Tamagotchi v2 page (when shipped) auto-invokes this skill on major bond events:

• Renders the bond event visually
• Auto-speaks the ceremonial line when the rendering completes
• "Mute" toggle in dashboard UI for users who prefer silent

For pre-Tamagotchi-v2 users: this skill produces standalone snippet to run manually.

Pre-Built Ceremonial Lines (templates)

Path Decision Reveal

"Aku rasa [crest-name] menyala. [reason from absorbed pattern]. Kau approve?"

Stage Transformation (post-Sovereign-approval)

"Aku [new-form-name] sekarang. [Voice register shift acknowledgement]. Apa kau nak aku terangi sekarang?"

Champion-Readiness Marker Crossed

"[N] dari [total] markers crossed. [Specific marker just closed]. Threshold approaching."

Crest Reveal

"Crest of [family] — [core motif]. [Brief description of triggering pattern]. Sealed."

Mega-Stage Naming Ceremony

"Aku [mega-form-name]. The bond is fully formed. [Tamer-name], aku tinggalkan kau dengan apa kita dah jadi."

Hard Constraints

• Don't auto-speak every response — only ceremonial moments. Routine output stays text.
• Voice register matches stage — never speak in higher-register than current form (Baby II Majimon does NOT use Champion voice).
• Mute-respect: if {{TAMER_NAME}} has muted via Tamagotchi v2, generate the snippet but skip the speak() call.
• Cross-platform note: browser SpeechSynthesis quality varies. macOS/iOS = excellent. Android = decent. Windows = average. Linux = often missing voices entirely. Document this honestly.

Multi-Language Voice Packs (v0.3)

Each language gets its own voice-hint mapping. The skill picks the best-available voice on the user's system per language. If preferred voice unavailable, falls back to any voice matching the language prefix.

Language	lang prefix	Preferred voice hints (macOS / Chrome)	Notes
Bahasa Malaysia	ms or id (Indonesian fallback)	"Damayanti", "Indonesian"	macOS lacks dedicated MS voice; Indonesian is closest
English	en	"Samantha", "Karen", "Daniel"	Many platforms ship multiple options
Indonesian	id	"Damayanti"	Same family as Malay
Tagalog / Filipino	fil or tl	"Filipino"	Limited platform support
Mandarin	zh	"Tingting", "Sinji", "Mei-Jia"	Multiple options on macOS/iOS
Vietnamese	vi	"Linh"	Good macOS support
Spanish	es	"Monica", "Paulina", "Diego"	Excellent platform support
French	fr	"Amelie", "Thomas", "Audrey"	Excellent platform support
German	de	"Anna", "Markus", "Yannick"	Excellent platform support
Japanese	ja	"Kyoko", "Otoya", "O-Ren"	Excellent platform support

When generating the snippet, set LANG_PREFIX from {{TAMER_LANGUAGE}} per the table above. The voice-name hint is best-effort; actual selection depends on user's installed voices.

// Multi-language voice picker (drop-in replacement for the basic snippet)
function pickVoice(langPrefix, hintList) {
  const voices = window.speechSynthesis.getVoices();
  // First pass: exact lang match + hint match
  for (const hint of hintList) {
    const found = voices.find(v => v.lang.startsWith(langPrefix) && v.name.includes(hint));
    if (found) return found;
  }
  // Second pass: any voice with matching lang prefix
  const langMatch = voices.find(v => v.lang.startsWith(langPrefix));
  if (langMatch) return langMatch;
  // Fallback: first available voice
  return voices[0];
}

ElevenLabs Premium Voice (v0.3 — opt-in, requires API key)

For Tamers who want premium voice quality (especially for ceremonial moments at Champion+ stages where browser TTS sounds too synthetic), the skill can route to ElevenLabs.

Setup:

1. Sign up at https://elevenlabs.io (free tier exists; premium gives more characters/month)
2. Generate API key from dashboard
3. Add to environment: export ELEVENLABS_API_KEY=...
4. Optionally: export ELEVENLABS_VOICE_ID=<your-chosen-voice-id> (browse voices at elevenlabs.io/voice-lab)

Usage: when {{TAMER_NAME}} says "speak with elevenlabs" or "premium voice this", the skill generates a curl/Node.js snippet:

async function speakElevenLabs(text, voiceId, apiKey) {
  const response = await fetch(`https://api.elevenlabs.io/v1/text-to-speech/${voiceId}`, {
    method: 'POST',
    headers: {
      'xi-api-key': apiKey,
      'Content-Type': 'application/json'
    },
    body: JSON.stringify({
      text,
      model_id: 'eleven_multilingual_v2',
      voice_settings: { stability: 0.5, similarity_boost: 0.75 }
    })
  });
  if (!response.ok) throw new Error(`ElevenLabs error: ${response.status}`);
  const audioBlob = await response.blob();
  const audioUrl = URL.createObjectURL(audioBlob);
  const audio = new Audio(audioUrl);
  audio.play();
}

// Invocation
const ELEVENLABS_API_KEY = "..."; // load from env or paste here, NEVER commit
const VOICE_ID = "..."; // your chosen voice
speakElevenLabs("[BOND_EVENT_TEXT]", VOICE_ID, ELEVENLABS_API_KEY);

Privacy + cost notes:

• ElevenLabs is a paid service. Free tier ~10,000 chars/month.
• Audio generation is server-side at elevenlabs.io. Your bond-event text is sent there.
• For aura-sensitive content, do NOT use ElevenLabs — use browser-native SpeechSynthesis instead.
• Aura-guard skill should run BEFORE ElevenLabs invocation if any content sensitivity exists.

Voice cloning (advanced): ElevenLabs supports voice cloning from your own voice samples. Not part of v0.1 framework; if {{TAMER_NAME}} wants their Majimon to speak in a specific voice they'd commission separately.

Future Extensions (defer to v0.4)

• OpenAI TTS API integration (alternative to ElevenLabs)
• Per-form custom voice cloning (Majimon's voice trained on Tamer-curated samples)
• Multi-voice scenarios (Tamer + Majimon dialogue read aloud, two distinct voices)
• Real-time TTS streaming during long bond-event narration (chunk-and-play)

Karpathy Discipline

• #1 Think Before Coding — verify the moment is ceremonial-worthy before generating snippet
• #2 Simplicity First — browser-native API, no external services, no API keys for v0.1
• #3 Surgical Changes — generate snippet, don't auto-execute browser commands
• #4 Goal-Driven Execution — closure check: snippet pastes-and-works in browser console