30 to 60 seconds is enough
Read the coached script for 30–60 seconds — up to three takes — and Fish Audio builds an accurate replica. No studio session required.
NARRATE
Voice cloning
Record a short sample and CourseSmith clones your voice with Fish Audio’s expressive S2.1 model — so every course sounds like you, with real intonation, without recording every line. Part of the Pro plan.
30–60 s sampleUp to 3 takesFish Audio S2.1Emotional prosodyPro
Powered by Fish Audio
Your voice, cloned — and expressive.
CourseSmith clones your voice with Fish Audio's S2.1 model — the same tech behind millions of expressive AI voices. Instead of a flat recording, you read a short, coached script that captures how you actually pace and stress a lesson, so the clone sounds like you, not a monotone.
STEADY AND DELIBERATE — COUNT OFF THE POINTS
“The first principle is simple: pause when you’d naturally pause. The second is harder: be willing to slow down on the parts that matter. And the third? Have fun with it.”
List cadence with a rising-falling-rising pattern. Numbered items are how courses actually sound; the clone needs this rhythm.
ANIMATED — EMPHASIZE “FAVORITE”
“Today I want to walk you through three ideas. First, how attention actually works. Second, why timing matters more than content. And third — my favorite — how to make a five-minute lesson feel like a conversation.”
Emphasis on a single word teaches the clone where you stress key terms.
CALMLY HANDING OFF
“Let’s start with the first one.”
A short ending sentence with falling intonation. Mirrors how courses transition between sections.
0:01
✓ Add to clone
A 30–60 second sample
Read the coached script for 30–60 seconds, in up to three takes — no studio session.
Expressive, not robotic
The S2.1 model captures emotion and prosody, so your clone carries real intonation.
A coached script
Read varied cadences — steady, animated, calm hand-offs — so the clone learns how you pace a lesson.
One voice, whole library
Every lesson, short and trailer can use your clone — one identity across everything you make.
What it does
The detail that makes it work.
Expressive, not robotic
The S2.1 model captures emotion and prosody, so your clone carries real intonation rather than a flat monotone.
A coached script for a better clone
CourseSmith hands you a short script with varied cadences — steady, animated, calm hand-offs — so the clone learns how you actually pace a lesson.
One voice for the whole library
Every lesson, short and trailer you make can use your clone, so your courses keep one identity.
Learn CourseSmith
Watch the short tutorial.
More in voice & language
Pairs well with these.
Build a course with voice cloning.
It’s one part of a tool that carries your source all the way to a polished, finished video.