Content Library (LinkedIn Growth Mastery)·LinkedIn Growth Mastery - AI Video Content Creation·42:40

Build Your AI Twin: Clone Your Face on HeyGen and Your Voice on ElevenLabs (Instant vs Professional Cloning)

Uttam Gupta Trainer - former director of growth at Growth School, founder of Agent Valley (agentvalley.ai)

The short version

HeyGen Creator: one cloned face, many 'looks' - record several outfits in one session. Signup, plan tiers (Free 1 video/month with watermark; Creator $29; ~$99 above), upload footage, verify on camera, 'your digital twin is ready' (0:00-0:20).
ElevenLabs (~$11 then ~$22/month): Instant cloning from 10-30 seconds sounds wrong (wrong accent live); Professional cloning needs 30+ minutes of clean single-speaker audio in the delivery style you will publish in, and is indistinguishable (0:20-0:39).
Generation settings: Multilingual v2 over v3 for cloned fidelity, Stability ~70, Similarity 70-80, regenerate on misses; CAPS for emphasis and break tags for pauses (0:36-0:42).

At a glance, three clicks deep

Skim here first: the closed row is the glance, open is the study card with the key points and timestamps, and the ↓ link drops to that concept's full write-up below.

01HeyGen: clone the face, multiply the looksClone real person -> verify -> looks per outfit;›

Clone real person -> verify -> looks per outfit; Design with AI for more.

Plan tiers and tour (0:03-0:10)

Upload and verify (0:10-0:14)

Design with AI (0:14-0:17)

↓ Full write-up of this concept

02ElevenLabs: instant vs professional voice cloningProfessional clone, 30+ min, matched style.›

Professional clone, 30+ min, matched style.

Instant vs professional (0:20-0:26)

Recording rules (0:23-0:30)

Side-by-side test (0:30-0:39)

↓ Full write-up of this concept

03Voice generation settings that workv2, Stability 70, Similarity 70-80, regenerate.›

v2, Stability 70, Similarity 70-80, regenerate.

Model and sliders (0:36-0:40)

WhatsApp micro-podcast (0:39-0:41)

↓ Full write-up of this concept

The concepts in full

HeyGen: clone the face, multiply the looks

One face per account - but every outfit you filmed is another look, and 'Design with AI' invents more.

PROCEDURE: HeyGen -> Avatars -> 'Clone a real person' (not virtual character) -> Upload Footage (file or Drive link) -> name -> processing -> identity verification (read a phrase on camera) -> twin ready. Creator plan holds one real face with unlimited looks, so record several outfits in one session; more people need a higher tier. 'Design with AI' re-renders the trained avatar into new settings (a podcast-mic setup) from a preset, paid tiers only. HeyGen's built-in voice clone is shown and dismissed as lower quality than ElevenLabs.

Why it matters

Variety without a second shoot.

This idea elsewherecontrastsOH XII argued for real-face video over avatars

ElevenLabs: instant vs professional voice cloning

Thirty seconds gives you a stranger with your name. Thirty minutes gives you you.

Instant Voice Cloning: 10-30 seconds, available on low tiers, noticeably off (the live test gave the wrong accent). Professional Voice Cloning: Creator plan, minimum 30 minutes of clean single-speaker audio (the trainer used over an hour, his colleague two), recorded in the energy and pace you will publish in - don't read a book slowly if the output is fast reels; upload in stages, then a verification read. Side by side on the same 'secret Japanese city / Woven City' script the professional clone was indistinguishable. Best practices: write naturally, CAPS for emphasis, '<break time="1s" />' for pauses; ElevenLabs' own best-practices article.

Why it matters

The single biggest realism lever in the pipeline.

This idea elsewhereparallelBasecamp 6 on ElevenLabs voices parallelBC9's ad pipeline uses ElevenLabs for voice

Voice generation settings that work

The newest model is not the best one for your clone.

Text-to-Speech: pick Multilingual v2 rather than v3 for cloned-voice fidelity (Flash 2.5 for speed); Stability around 70, Similarity 70-80; Regenerate when a take mispronounces. Aside: the trainer uses this to publish a daily AI-news voice note to a WhatsApp community - a 'micro podcast' from text.

Why it matters

Defaults are not production settings.

Tools referenced

Tool	Coverage	Moment	Context
HeyGen	demonstrated		Signup, avatar, verification, Design with AI; $29 / ~$99
ElevenLabs	demonstrated		Instant vs professional cloning; TTS settings; ~$11/$22

Action items

Extraction notes

This page was built from an auto-generated transcript, which garbles product and people's names. Those were corrected silently in everything above and logged here for transparency. The warnings flag claims that were true on the recording day but change fast.

Transcript corrections applied

The transcript says	The trainer actually means
Lignin / lean in	LinkedIn
GrowthCool / GrowSchool	Growth School
agent value	Agent Valley
Sutta / Utham / Autumn	Uttam (trainer)

True on recording day — verify before relying

HeyGen and ElevenLabs plan prices, plan limits and model names (Multilingual v2 / v3, Flash 2.5) are recording-day details
The course-provisioned HeyGen plan is an Outskill bundle that may not persist