Voice Cloning with ElevenLabs

ElevenLabs is the market leader in AI voice generation. If you’re going to use AI voice for your courses, this is the tool to learn. Here’s the complete workflow.

Step 1: Record Your Voice Sample

Voice cloning starts with a sample of your real voice. The quality of this sample directly determines the quality of your cloned voice.

Minimum sample: 3 minutes of clean, clear recording. Recommended sample: 30 minutes for best results.

How to Record the Sample

1. Use the same microphone and room setup you’d use for actual recording
2. Read something you’re comfortable with — your course outline, a blog post, a book chapter
3. Speak naturally, at your normal pace and energy level
4. Don’t act or perform — just talk as you would in a real lesson
5. Avoid long pauses, heavy breathing, or background noise

What to read: Your existing course content is ideal. Read several lesson scripts or outlines. The more varied the content (questions, explanations, lists, stories), the better the clone will handle different styles.

File format: Export as WAV or high-quality MP3 (320kbps). ElevenLabs accepts both.

Step 2: Upload and Clone

1. Create an account at elevenlabs.io
2. Go to Voices → Add Voice → Instant Voice Cloning
3. Upload your audio file(s)
4. Name your voice (e.g., “My Course Voice”)
5. Add a brief description (e.g., “Warm, clear, educational tone”)
6. Click Add Voice

Processing takes a few minutes. Once complete, you can type any text and hear it in your cloned voice.

Step 3: Test and Iterate

Type a paragraph from one of your course lessons and generate the audio. Listen critically:

• Does it sound like you? The clone should be recognizable as your voice.
• Is the pacing natural? If it rushes through sentences, add commas and periods to your script to force pauses.
• Are technical terms correct? AI often mispronounces domain-specific words.

Generate several tests with different content types (a story, a list, a question-and-answer) to see how the clone handles variety.

Step 4: Fine-Tune Pronunciation

AI voice tools struggle with:

• Technical terms and jargon
• Brand names and product names
• Acronyms (should “GHL” be “G-H-L” or “gull”?)
• Foreign words and names

ElevenLabs pronunciation dictionary: You can add custom pronunciations for words the AI gets wrong. Go to Voices → Pronunciation Dictionary and add your terms with the correct phonetic spelling.

Workaround: In your scripts, spell out how you want things pronounced. Instead of “GoHighLevel,” write “Go High Level.” Instead of “ATR-2100x,” write “A-T-R twenty-one hundred X.”

Step 5: Generate Your Audio

Once your voice is cloned and tested:

1. Paste your lesson script into the ElevenLabs text editor
2. Select your cloned voice
3. Adjust Stability and Clarity + Similarity Enhancement:
   • Stability (low = 30–50%): More expressive, more varied. Better for stories and motivation.
   • Stability (high = 60–80%): More consistent, more controlled. Better for instructions and technical content.
   • Clarity + Similarity (high = 75–90%): Sounds more like your original voice.
4. Click Generate
5. Download the audio as MP3

Batch generation: You can paste an entire lesson script and generate it at once. For long lessons, generate in sections (5–10 paragraphs at a time) for better quality control.

Step 6: Edit the Generated Audio

AI-generated audio still needs editing:

• Cut awkward pauses where the AI hesitated
• Remove mispronunciations (re-generate just that section)
• Normalize the volume to match your real-voice recordings (if using the hybrid approach)
• Add intro/outro music and transitions (covered in Lesson 11)

Import the AI-generated MP3 into Audacity alongside any real-voice recordings and treat it like any other audio — cut, level, and export.

Cost Planning

ElevenLabs pricing is per character:

Plan	Characters/month	Cost	Good For
Free	10,000	$0	Testing
Starter	30,000	$5/mo	1–2 short courses
Creator	100,000	$11/mo	Regular course production
Pro	500,000	$22/mo	Multiple courses + marketing content

Rough math: A 12-lesson course with ~1,200 words per lesson = ~72,000 characters. That’s about one course per month on the Creator plan.

Your Action Step

If AI voice interests you: record a 5-minute voice sample, upload it to ElevenLabs, and clone your voice. Generate 60 seconds of course content. Listen to the result. This test costs nothing on the free tier and tells you everything you need to know about whether AI voice works for your courses.

Next up: the audio editing workflow — whether you recorded yourself or generated with AI.