🔊 Clone your voice with AI (and control the emotion)

Your voice is the most personal asset your content has — and AI voice platforms (Fish Audio, ElevenLabs, and similar) can now clone it well enough to narrate content you never sat down to record. Here's the workflow, including emotion control and the consent rules that aren't optional.

Step 1: Record clean source audio

• A few minutes of you talking naturally — conversational, not "announcer voice." Most platforms want 1–10 minutes.
• Quiet room, phone or mic 6–8 inches away, no music or background noise.
• Vary it: some energetic lines, some calm ones. The model learns your range from the sample.

Step 2: Create and test the clone

Upload to your platform, generate a test paragraph, and listen for the two common failures: flattened energy (fix: re-record livelier source) and mushy consonants (fix: cleaner audio, closer mic). Iterate the source until the test sounds like you on a good day.

Step 3: Control the emotion

The difference between robotic and real is direction. Two levers:

• Platform controls: most tools expose energy/stability/pacing settings, and some support per-line emotion tags. Learn yours.
• Script-side direction with Claude:

Where a cloned voice earns its keep

Narrating faceless video content, ad variation batches, turning written posts into audio, and multilingual versions of your content in your own voice. Anywhere your voice adds trust but recording time is the bottleneck.