ElevenLabs Integration

Turn text into realistic speech with AI

What is ElevenLabs?

ElevenLabs creates incredibly realistic AI voices. You can turn any text into natural-sounding speech - perfect for audiobooks, accessibility, voice assistants, and more.

Perfect for:

• Text-to-speech for accessibility
• Audiobook creation
• Voice assistants
• Content narration
• Podcast generation

What You’ll Need

1. ElevenLabs Account

   Sign up at elevenlabs.io

2. Choose Voice

   Pick from their voice library

3. API Key

   Get your API key

4. 5 Minutes

   Quick setup!

Step 1: Get Your ElevenLabs API Key

Create Account

1. Go to elevenlabs.io
2. Click Sign Up
3. Verify your email

Free Plan:

• 10,000 characters per month
• All voices available
• Good for testing!

Paid Plans:

• Starter: $5/month (30,000 characters)
• Creator: $22/month (100,000 characters)
• Pro: $99/month (500,000 characters)

Get Your API Key

1. Click your Profile icon (top right)
2. Select Profile + API Key
3. Copy your API key

Choose a Voice

1. Go to Voices in the dashboard
2. Listen to different voices
3. Click on one you like
4. Copy the Voice ID

Popular Voices:

• Rachel - Calm, friendly female voice
• Adam - Deep, authoritative male voice
• Bella - Soft, expressive female voice
• Josh - Professional, clear male voice

Step 2: Tell the Agent What You Want

Describe what you’re building:

Example 1: Article Reader

I want to convert blog posts to audio.

ElevenLabs API key: [paste your key]
Voice: Rachel (or Voice ID: xyz)

Features:
- "Listen" button on each article
- Generates audio from the text
- Audio player to listen
- Download audio option

Please set this up.

Example 2: Audiobook Generator

I need to create audiobooks from text.

ElevenLabs API key: [paste your key]
Voice: Adam

Features:
- Upload a text file or paste text
- Convert entire book to audio
- Chapter-by-chapter playback
- Download as MP3

Please build this.

Example 3: Voice Assistant

I want an AI that speaks responses.

ElevenLabs API key: [paste your key]
Voice: Bella

Flow:
1. User types a question
2. AI generates text response
3. Convert response to speech
4. Play audio automatically

Please set this up.

That’s it! The agent will handle all the voice generation setup.

Step 3: Test Your Voice

After setup:

1. Click Preview

   In Emergent, click Preview

2. Enter Text

   Type or paste some text

3. Generate Speech

   Click the generate/convert button

4. Listen

   Should hear realistic AI voice!

5. Try Different Text

   Test with various content

6. Check Quality

   Make sure it sounds natural

Common Questions

How much does it cost?

Based on characters converted:

Free: 10,000 characters/month

• About 1,500 words
• Great for testing

Paid: Starting at $5/month for 30,000 characters

1,000 words = roughly 6,700 characters

Can users choose different voices?

Yes! Tell the agent:

Let users pick from multiple voices:
- Rachel (female, friendly)
- Adam (male, professional)
- Bella (female, warm)

Add a voice selector dropdown.

Speech sounds robotic. Can you make it better?

ElevenLabs is already very natural, but you can tell the agent:

The voice sounds a bit flat. Can you:
1. Use a more expressive voice
2. Adjust speech settings for more emotion
3. Add pauses at punctuation

Can I create a custom voice?

Yes, with voice cloning! Tell the agent:

I want to clone a specific voice.
Users will upload audio samples.
Generate speech using that cloned voice.

Please add voice cloning feature.

(Requires paid ElevenLabs plan)

Something Not Working?

Audio Not Generating

Tell the agent:

When I click generate, nothing happens.
Can you check the ElevenLabs connection?

Poor Audio Quality

Tell the agent:

The audio quality is poor or choppy.
Can you:
1. Use higher quality settings
2. Check the audio encoding

Generation Too Slow

Tell the agent:

It takes too long to generate speech.

Can you:
1. Add a loading indicator
2. Show progress
3. Optimize the generation process

Hitting Character Limits

Tell the agent:

I'm running out of characters.

Can you:
1. Show how many characters used
2. Warn when close to limit
3. Split long text into chunks

Real Examples

Example 1: Learning Platform

I'm building an e-learning platform.

Features:
- Convert lesson text to audio
- Students can listen while reading
- Downloadable lesson audio
- Multiple voice options

ElevenLabs API key: [paste key]

Please build this.

Example 2: News Reader

Create an audio news reader.

Features:
- Fetch daily news articles
- Convert to audio automatically
- Play in podcast-style player
- Generate new episodes daily

ElevenLabs API key: [paste key]
Voice: Professional male voice

Please set this up.

Example 3: Accessibility Tool

Make my website accessible with audio.

Features:
- "Listen" button on all pages
- Reads page content aloud
- Adjustable speed
- Pause/resume controls

ElevenLabs API key: [paste key]

Please build this.

Advanced Features

Once basic speech works:

Multiple Languages

Add support for multiple languages:
- English
- Spanish  
- French
- German

Auto-detect language and use appropriate voice.

Emotion Control

Add emotion options:
- Happy/Excited
- Calm/Professional
- Serious/Formal

Let users choose the tone.

Speech Customization

Let users adjust:
- Speech speed (0.5x to 2x)
- Voice pitch
- Emphasis on certain words

Quick Checklist

Before going live:

• Tested with short text
• Tested with long text
• Audio quality is good
• Tried different voices
• Works on mobile
• Download option works (if needed)
• Monitored character usage

Remember

Free tier for testing - 10,000 characters/month

Very natural voices - ElevenLabs has the best quality

Monitor usage - Check your character count

Test with real content - Try actual articles/text

Stuck? Tell the agent: “I need help with ElevenLabs voice generation” and describe the issue.