tts-content

name	tts-content
description	Generate speech audio using Gemini TTS. Use when user asks for "text to speech", "voiceover", "narration", "voice", "audio", "podcast", "dialogue", "read this", "speak this", or any speech synthesis request.
license	MIT
allowed-tools	["generateVoiceover","generateMultiSpeakerVoiceover"]

Text-to-Speech Content Skill

When This Skill Activates

• User wants text converted to speech
• User needs voiceover for video
• User wants podcast/dialogue audio
• User asks for narration

---

STEP 1: Ask Audio Type

Ask the user:

What type of audio do you need?

1. Single Narrator - One voice reading text
2. Multi-Speaker - Dialogue, podcast, conversation
3. Voiceover - For video/presentation
4. Character Voice - Specific personality/style

Store answer as: audioType

---

STEP 2: Ask Voice Preference

For Single Narrator:

What voice characteristics do you want?

Gender:

1. Female voice
2. Male voice

Tone:

1. Professional/Authoritative
2. Friendly/Warm
3. Energetic/Excited
4. Calm/Soothing
5. Casual/Conversational

For Multi-Speaker:

How many speakers do you need? (2-4)

For each speaker, I'll ask about their voice.

---

STEP 3: Voice Selection

Female Voices

Voice	Characteristics	Best For
Zephyr	Warm, friendly, versatile	General narration
Aoede	Bright, energetic	Excited content
Kore	Calm, soothing	Meditation, ASMR
Leda	Professional, clear	Corporate, tutorials
Achernar	Authoritative	News, announcements
Pulcherrima	Elegant, refined	Luxury, premium

Male Voices

Voice	Characteristics	Best For
Puck	Friendly, casual	Casual content
Charon	Deep, authoritative	Documentary
Orus	Warm, trustworthy	Explainers
Fenrir	Energetic, dynamic	Ads, promos
Schedar	Professional, clear	Corporate
Achird	Calm, measured	Tutorials

Voice Recommendation by Content Type

Content Type	Recommended Voices
Tutorial/Explainer	Leda, Achird, Orus
Podcast	Zephyr + Puck (duo)
Advertisement	Fenrir, Aoede
Meditation/Calm	Kore, Charon
Corporate	Achernar, Schedar
Storytelling	Zephyr, Orus

---

STEP 4: Ask Style Settings

Ask the user:

How should the voice sound?

Style:

• neutral - Standard delivery
• vocal_smile - Bright, sunny
• newscaster - Professional news
• whisper - Soft, intimate
• excited - High energy
• casual - Relaxed, informal
• serious - Grave, formal

Pace:

• natural - Normal speed
• slow - Deliberate, measured
• energetic - Quick, upbeat
• rapid_fire - Very fast

Accent (optional):

• american_general - Standard US
• british_rp - British received pronunciation
• australian - Australian accent

Store answers as: style, pace, accent

---

STEP 5: Format Text

For Multi-Speaker

Format the text with speaker labels:

Host: Welcome to the show!
Guest: Thanks for having me!
Host: So tell us about your project.
Guest: Well, it started when...

Audio Tags

User can add inline tags for expression:

[excited] This is amazing!
[whispers] But keep it a secret.
[laughs] That's hilarious!
[sighs] It's been a long day.
[slow] Listen carefully now.

Available Tags: [excited], [whispers], [laughs], [sighs], [gasp], [sarcastic], [serious], [slow], [fast], [short pause], [long pause]

---

STEP 6: Confirm and Generate

Show summary:

📋 TTS GENERATION PLAN
======================
Type: {audioType}
Voice(s): {voiceNames}
Style: {style}
Pace: {pace}
Text length: {wordCount} words

Estimated Cost: ~${cost}

Proceed? (yes/no)

Single Speaker Generation

Use the pre-built workflow function — it validates, retries, saves the .wav, and reports cost:

import { generateVoiceover } from '../index';

const result = await generateVoiceover({
  script: text,
  outputPath: 'output.wav',
  voiceName,
  voiceStyle: { style, pace, accent }
});
// result.data.audioPath, result.data.cost.totalCost

Multi-Speaker Generation

import { generateMultiSpeakerVoiceover } from '../index';

const result = await generateMultiSpeakerVoiceover({
  script: dialogueText,
  speakers: [
    { speaker: 'Host', voiceName: 'Zephyr', voiceStyle: { style: 'vocal_smile' } },
    { speaker: 'Guest', voiceName: 'Puck', voiceStyle: { style: 'casual' } }
  ],
  outputPath: 'podcast.wav'
});

---

Content Presets

Podcast Preset

speakers:
  - name: Host
    voice: Zephyr
    style: vocal_smile
    pace: natural
  - name: Guest
    voice: Puck
    style: casual
    pace: natural

Tutorial Preset

voice: Leda
style: neutral
pace: slow
accent: american_general

Advertisement Preset

voice: Fenrir
style: promo_hype
pace: energetic

Meditation Preset

voice: Kore
style: whisper
pace: slow
# Or use custom audioProfile:
audioProfile: "Calm, soothing meditation guide with gentle pauses"

---

All Available Voices

Female (14)

Achernar, Aoede, Autonoe, Callirrhoe, Despina, Erinome, Gacrux, Kore, Laomedeia, Leda, Pulcherrima, Sulafat, Vindemiatrix, Zephyr

Male (16)

Achird, Algenib, Algieba, Alnilam, Charon, Enceladus, Fenrir, Iapetus, Orus, Puck, Rasalgethi, Sadachbia, Sadaltager, Schedar, Umbriel, Zubenelgenubi

---

Cost Reference

Model	Input	Audio Output
gemini-2.5-flash-tts	$0.50/M tokens	$12.00/M tokens
gemini-3.1-flash-tts	$0.75/M tokens	$12.00/M tokens

Estimate: ~$0.001 per sentence, ~$0.02 per paragraph