CLONE ANY VIRAL [REEL].
ANY NICHE. YOUR FACE.

One Claude Code prompt. Drop a selfie. Drop a reel link. Get a finished 9:16 / 1080p / 60fps vertical video with your voice, your wardrobe, your niche — copying the reference's hook, shot grammar, edit rhythm and trending audio. Module 1 of the printer/ai content stack. Free.

Most "AI reel" apps make a stitched stock-footage slideshow with a robot voiceover that smells like a robot voiceover. This one does what an actual editor + actor + animator would do — locks your face across multi-angle shots, holds your identity through every scene, splits the trending instrumental off the source, mounts captions behind your head, and finishes in 1080p/60. The catch: you need a Higgsfield account (~$1.50 in credits per finished reel) and 5 minutes of one-time setup. That's it.

Get the prompt no email gate · just type your niche in chat · file URL drops in 30 sec
~$1.50
per finished reel · Higgsfield credits at fast/480p + 1080p60 upscale
28 min
from link to finished file · most of it is generation, not your time
7 modules
in the full content stack · this is module 01 · the warm-up layer

// tldr

You paste the prompt into a fresh Claude Code session as your first message. It onboards itself — checks tools, asks for keys, asks for your niche/CTA/voice, asks for one photo. Builds a multi-angle character board of you that every later video locks onto. From that moment on, the loop is: you drop a reel link → it ships a 9:16/1080p60 video starring you, in your niche, in 28 minutes. Same hook device as the reference, your words, your face, the reference's trending audio reattached. The prompt itself is one .md file. The agent below will hand you the download URL after a quick exchange.

FOUR STEPS. NO [EDITOR]. NO ACTOR.

  1. 01

    One-time setup · 5 min

    Paste the prompt as your first message in a fresh Claude Code session. It checks your tools (ffmpeg, python, Higgsfield CLI), asks for the keys it needs once (Higgsfield, Gemini for video breakdown, Apify for scraping), writes them to .env, and runs a cheap green-light test on each one. After this you never touch it again.

  2. 02

    Hero board · 1 photo → multi-angle character

    Drop one clean selfie. The prompt generates your character board: front / 3⁄4 / side / back full-body + front / 3⁄4 / profile face headshot. Identity locked. Every later video clip references this board, so the operator on screen never drifts between scenes. No photoshoot. One selfie.

  3. 03

    Onboarding brief · once, then never again

    One short brief — your niche, your CTA codeword, voice preference (your real cloned voice via ElevenLabs, or Seedance's native voice that approximates yours), default wardrobe + setting. Saved to brief.md and reused for every reel after.

  4. 04

    Per-link loop · paste link → get reel

    You paste a viral reel URL. Apify scrapes the source video + caption locally. Gemini breaks it down — verbatim speech with timestamps, every on-screen text overlay, the shot list, the edit rhythm, the structure. The prompt rewrites the script for your niche, mirroring the reference beat-for-beat. Seedance 2.0 generates every clip (you on camera + dynamic b-roll). ffmpeg stitches the cut, mounts the captions behind your head, ducks audio, reattaches the trending instrumental. Bytedance upscale finishes 1080p / 60fps. Done.

THE MATH VS [EVERY OTHER PATH].

You wanted reels that look like the ones eating right now. Until this prompt existed, you had four bad options. Here's what each one actually costs once you do the math nobody likes to write down.

Option $ / reel Time / reel Your face? Trending audio?
Hire an editor $200–500 3–7 days no — stock no
Buy an "AI reel" app $99/mo + 2–4 hrs your time 2–4 hrs sometimes rarely
Hire a UGC creator $300–800 5–10 days their face no
Do it yourself · iPhone + CapCut $0 + 4–8 hrs 4–8 hrs yes yes
The Reel Clone Machine ~$1.50 ~28 min yes · identity-locked yes · auto-attached

The "do it yourself" line is the one most operators don't price honestly. If your hourly rate is $50, "free" reels cost $200–400 each. The reel clone machine drops that to ~$1.50 and gives the time back.

ONE PROMPT. EIGHT SECTIONS. [NO BABYSITTING].

The .md file is a Claude Code system prompt — you paste it once as your first message and Claude becomes a content operator inside your filesystem. It runs proactively, asks only when a decision is genuinely yours, and refuses to ship a clip that drifted off your face. Here's the skeleton.

§ 0

Identity & boundaries

What the machine does, where the line is (replication, not theft — formats copied, scripts rewritten, only your face ever generated).

§ 1

First-run setup

Tool detection (ffmpeg, python, rembg, Higgsfield CLI), key collection from .env, one cheap test per service, then writes ./.studio/ready.

§ 2

Hero board · one selfie

Image-to-image identity pass → multi-angle character contact sheet → wardrobe board. Saved to ./hero/. Approved before any video runs.

§ 3

Onboarding brief

Niche, offer, CTA codeword, adaptation depth (tight/balanced/loose), voice strategy, language, look & setting, music source. Saved to brief.md.

§ 4

Per-link loop

Scrape → break down (Gemini) → script as a beat table → generate clips (Seedance 2.0 fast/480p) → assemble (ffmpeg, captions-behind-subject, trending instrumental) → QC (Gemini) → finish (1080p/60).

§ 5

Locked production rules

The non-negotiables: 480p generation + 1080p60 upscale, voice recipe (12s clean ref → native Seedance speech, no "clone" word), dynamic > static head, captions above the head-line, trend audio preserved with demucs.

§ 6

Delivery report

What you see at the end: file path, what format was cloned, exact lines spoken, ready-to-paste caption + CTA, audio reattach note, credit spend.

§ 7

Defaults & etiquette

Money discipline (~60–90 credits per reel; warns before spending more), regenerates failed clips instead of shipping them, keeps a running log.md of every link → slug → credits.

ONE FORMAT > ZERO. [SEVEN BEATS ONE].

The Reel Clone Machine is the warm-up layer of the printer/ai content stack. Run it solo and you'll punch above your weight in any niche where reels carry weight — parenting, fitness, finance, real estate, beauty, B2B founder content. We've seen it pull $1–3k/mo on its own for operators with a clean offer underneath.

That's also the ceiling. The reason: one format hitting one slot of the algorithm. Reels are the warm-up. The other six modules in the stack hit different slots — scripted talking-head reels, AI-cinematic narrative shorts, podcast-clip pipelines, hooked carousel posts, long-form-cut Shorts, live-Zoom hooks. Same audience, different formats stacked on top of each other. The math compounds 3–10× because the algorithm stops betting on one horse and starts feeding you across the whole content surface.

M_01
Reel Clone Machine ← you are here
Cloned vertical reels from any reference link.
M_02
Scripted Creator Reels
Talking-head reels with a written hook bank + scripted-creator persona.
M_03
AI-Cinematic Shorts
Narrative shorts with cinema-grade lighting, blocking, depth — when reels need to feel like a movie.
M_04
Podcast-Clip Pipeline
Drop a podcast file. Get 10 viral clips with hook overlays, captions, b-roll auto-cut.
M_05
Hooked Carousels
Same hook math, static format. For the days the algorithm wants stills.
M_06
Long-Form Cut Shorts
Slice YouTube long-form into 5–15 vertical clips per video. Compounds discovery.
M_07
Live-Zoom Hooks
Hook overlays auto-detected from live calls / Zoom recordings → viral clips before the call ends.

The Reel Clone Machine alone is real, useful, and worth shipping. If it's all you ever use, you'll still be ahead of every operator hiring editors and UGC creators. If you want the compound effect — 3–10× the same math on the same audience — module 1 is where the door opens to the other six.

QUESTIONS PEOPLE [ASK] BEFORE THEY TAKE IT.

Most of these come up on every call. If your question isn't here, the agent in the corner will answer it — type it in.

How do I copy a viral Reel or TikTok with AI?

The Reel Clone Machine handles the whole pipeline. It scrapes the source video, breaks down its shot list and trending audio with Gemini, writes an adapted script in your niche, generates vertical clips with your face using Seedance 2.0, mounts captions behind your head, reattaches the trending instrumental, and finishes 1080p / 60fps. You paste one link, the agent ships a finished reel.

Is there a free AI tool to copy Instagram Reels or TikTok videos?

Yes — the Reel Clone Machine prompt is free. The runtime cost is about $1.50 per finished reel in Higgsfield credits (Seedance generation + bytedance upscale). There is no subscription and no checkout to get the prompt itself. The widget hands you the download URL after a 30-second exchange.

Can I clone a viral reel with my own face on camera?

Yes. You drop one selfie at setup. The prompt builds a multi-angle character board (front, 3/4, side, profile) that every later video clip references, so the operator on screen is consistently you — not a stock-footage stand-in. Identity holds across hook, b-roll, and close.

How long does it take to make an AI reel from a reference link?

About 28 minutes per reel. Most of it is the Seedance generation pass running unattended — your hands-on time is roughly 2-3 minutes (paste link, approve script, run finisher). You drop the link, walk away, come back to a finished 9:16 / 1080p / 60fps vertical file.

Will the Reel Clone Machine work for my niche?

Tested across fitness, finance, parenting, real estate, beauty, founder / B2B content, and creator niches. The prompt only copies the format and structure — the script is rewritten in your niche's voice. Any niche where short-form video has traction will work, which is most of them once you stop and think about it.

Do I need to know how to code to use this AI reel cloner?

No. You paste the file as your first message in a fresh Claude Code session. The agent onboards itself, asks for the keys it needs (Higgsfield, Gemini, Apify), and runs the whole loop conversationally from then on. Setup is 5 minutes, one time, then you never touch it again.

What about copyright — is this legal?

The prompt copies format and structure, not the script verbatim. The reference's hook device is niche-adapted; the body is rewritten in your words. The only human likeness ever generated is yours, from your own photo. No reference footage is reused.

Where does this fit in the printer/ai system?

It's module 1 of 7 — the warm-up content layer. The other six (scripted-creator reels, ai-cinematic shorts, podcast-clip pipeline, hooked carousels, long-form-cut shorts, live-Zoom hooks) hit different algorithm slots so your content surface compounds 3–10× rather than betting one horse. See the full system →

I built the first version of this prompt because I was paying an editor $400 per reel and getting back something a TikTok mom could outpace with a ring light. The honest version: the file you're about to get is the same prompt I run on my own accounts. I drop a reel link in the morning, walk away, come back to a finished vertical video that looks like the reference's exact format with my face in it. Repeat across however many references I have queued.

The reason it works: it doesn't pretend AI video is good enough to generate from a vibe. It takes a real successful reel, has Gemini study what makes it work, and tells Seedance "make me that shot, with this face, in this niche". The reference is the prior. Your operator board is the constraint. The output is something that wouldn't exist without both.

Take it. Use it. If it earns its keep, the rest of the stack is at printerai.dev.

— DROP, M_01 operator

PASTE. DROP A LINK.
WATCH IT [SHIP].

The file is yours below — no email gate, no checkout. The agent on screen will hand you the download URL after a 30-second exchange (it asks for your niche so it can tell you which reel formats are eating in your space right now — useful info regardless of whether you take the prompt or not).

you'll need a Higgsfield account (~$1.50 in credits per reel) and 5 minutes for one-time setup. that's the whole catch.

— DROP