DEV Community

Ehmad from Qlue Dental
Ehmad from Qlue Dental

Posted on AI-assisted

How to make 30 faceless YouTube videos in 30 minutes using Claude

At 21, Noah Morris was making about $1.8 million a year from around 20 faceless YouTube channels, by his own account in interviews.

Two years earlier the headline had been $30,000 a month.

Then a copyright claim over one short audio clip got a channel terminated, and every channel linked to it went down with it. Close to $50,000 a month in combined revenue, gone, as he told NetInfluencer. Hold onto that. We will come back to it.

He is the extreme. The normal version, from Korpi's 2026 ranges:

  • 1,000 to 10,000 subscribers: $50 to 500 a month
  • 10,000 to 50,000: $500 to 3,000
  • 50,000 to 100,000: $2,000 to 8,000
  • 100,000 to 500,000: $5,000 to 25,000
  • 500,000 and up: $15,000 to 80,000 and more

And the top of the category, as estimated by GoFaceless: Bright Side at $80,000 to 250,000 a month.

Daily Dose of Internet: around $270,000 a month.

Two things those numbers hide.

First, they are monthly. Monthly means a catalog. Nobody earns $5,000 a month from one video. They earn it from 80 videos that each bring a little.

Second, the people at the top run channels in the plural. Noah had 20.

So the real skill is not making a video. It is making a lot of videos that are each worth watching. That is a pipeline problem, and this article is the pipeline.

What a faceless channel is, in one paragraph

Nobody appears on camera. The video is a narrator over pictures, maps, animation or diagrams. History explainers, science, true crime, finance, sleep stories, "what if" videos. The audience follows the format, not a person. That is why it can be built fast, run in plural, and sold.

What "30 videos in 30 minutes" means here

Thirty films queued in about 30 minutes of your own attention. Roughly one minute per film: paste the brief, press Generate, read the script, approve. The films are then made on a server while you do something else.

I have not timed 30 in a row, and every script you edit takes longer. It is a budget, not a promise.

The two jobs

Here is the thing most people get wrong. They try to think and produce at the same time, in the same tool, in the same sitting. Ideas, facts, script, pictures, voice, edit. Everything comes out average.

Split it.

  • Claude does the thinking: the niche, 30 ideas, the facts to check, a story brief for each.
  • A video tool does the production: the script from your brief, the pictures, the voice, the title, the thumbnails.

You sit in the middle and read. That is the whole system.

Pick the video tool

There are several tools that turn a brief into a long faceless video now. Crreo, InVideo AI, Pictory and Fliki are the ones that come up most. I have looked at most of them.

I use Ziztu's AI YouTube Video Maker, because of four things that matter when you are making 30:

  1. It stops once, at the finished script, so you read before it spends.
  2. One look carries every picture, so film 27 looks like film 1.
  3. The price is on the button before each film is made.
  4. Real people appear as their real photo when the look is photographic, never an AI face.

The steps below are in Ziztu. The prompts work with any tool.

Step 1: Ask Claude for 30 ideas you can check

Open a folder for the channel and start Claude Code in it, or use a normal Claude chat. Paste this:

You are my YouTube research partner. My channel is a faceless
"history of engineering" channel. Videos are 8 minutes long.

Give me 30 video ideas as a table with these columns:
1. Working title
2. The one true event the video is about, with a date and a place
3. The hook: who, when and where in two sentences
4. The single question the video answers
5. One source I can check (a museum, a book, an encyclopedia page)

Rules: only events you are certain about. Mark any you are not
sure of with "?". Never put two ideas from the same era next to
each other. Save the table to ideas.md.
Enter fullscreen mode Exit fullscreen mode

Five rows in the shape it should come back in. These are facts from a public history-ideas list, so use them as a format check and verify every date for your own ideas:

# Working title True event Hook idea
1 Why the Pantheon dome never fell Pantheon, Rome, finished around AD 125 under Hadrian A dome 43 meters wide with no steel inside
2 The gear machine from a shipwreck Antikythera mechanism, found in a wreck off Antikythera in 1901 A bronze computer from the 2nd century BC
3 The stone that unlocked hieroglyphs Rosetta Stone, found 1799, carved 196 BC Three scripts, one message
4 How a baker's shop burned London Great Fire of London, from Pudding Lane, 2 September 1666 A fire that burned for about four days
5 Digging a canal between two oceans Panama Canal, opened 15 August 1914 A ditch that moved the sea

Do not skip the checking. Claude is excellent at ideas and sometimes wrong about dates. The "source I can check" column exists so you spend two minutes verifying instead of publishing a mistake.

Step 2: Custom mode, where Claude writes the brief and the tool makes the film

This is what makes each video different from the last. Instead of typing one line into the video tool, you give it a real brief.

Pick idea #1 and ask Claude:

For idea #1, write a story brief for an 8-minute film.

- 4 chapters. Each chapter has a goal and ends on an open question.
- The facts to include: numbers, dates, names, each with a source.
- Plain chronological order. No flashbacks, no gimmicks.
- Tone: warm storyteller.
- Which chapters are about a place, so the film can show a map.
- What to leave out.
Under 400 words. Save it as brief-01.md.
Enter fullscreen mode Exit fullscreen mode

Now open the AI YouTube Video Maker.

The AI YouTube Video Maker start screen with a box that takes a link or an idea, a Look picker, a Length picker and a Make my video button

The start screen. One box takes a link or an idea.

  1. Paste the brief into the box. Short briefs go straight in. If yours is long, paste the first lines and keep the rest for the Plan screen, where you can change any line.
  2. Set the length to 8 minutes. Any whole number from 3 to 60 works.
  3. Press Generate. You get a setup sheet for the look, the tone, the narrator and Basic or Pro, with the estimate in credits.
  4. Read the script. The tool stops once, at the finished script. You read it, change any line, and check the originality score. No picture is bought before you press "Make the film."

Pick the look so the channel feels like one place. There are 18, and one look carries every picture in the film:

The 18 looks: Map and cast, Old masters, Engraving, Cinematic real, Anime documentary, Paper cutout, Bedtime history, Cartoon comedy, Geometric Explainer, 3D Reconstruction, Archive Collage, Clay Stop-Motion, Brick World, Miniature World, Deep Space, Blueprint Cutaway, Dark Tech and Manga Ink

For an engineering history channel, Map and cast, Engraving or 3D Reconstruction fit well.

Then pick a narrator. There are 33, each with a short description and a sample you can play before you choose.

Step 3: The copy way, where you borrow a structure and tell your own story

The second way in is to paste the link of a YouTube video that is already working. The tool does not copy it. It studies how it is built: the title pattern, the hook, the chapters and the pace. Then it writes your version, about your topic, in your words.

Two safeguards keep it honest:

  • An originality score shows how far your script is from the source and marks the passages that sit too close. One click rewrites them.
  • Your own guide character. When a film copies a pasted video's drawing style, it gets its own simple guide figure, not the channel's mascot.

Five pictures drawn in the style of a pasted history video: a throne room, a guide next to a portrait, a winter river crossing, a second throne room and a relief map of Russia

A film drawn in the style of a pasted history video, with its own guide and a relief map.

A link only works when the video's captions can be read. If they cannot, paste the idea instead.

A good rhythm: use the copy way for the structure of your first video, then Custom mode with Claude's briefs for the next ones, so the channel builds its own voice.

Step 4: Queue the other 29

Ask Claude for the rest in one go:

Write briefs for ideas 2 to 30 in the same format. One file each:
brief-02.md, brief-03.md and so on. Vary the chapter structure so
no two videos open the same way. Keep a facts-to-check list at
the end of each file.
Enter fullscreen mode Exit fullscreen mode

Then work through them one at a time: paste, Generate, read the script, approve. This is where the 30-minute budget goes, at about a minute per film.

Three rules:

  1. Start with three, not thirty. Make three films and watch the price and the quality before you queue the rest. The price in credits is on the button before each film is made.
  2. Vary the looks and narrators across the batch. Do not upload 30 near-identical videos in one day.
  3. Read every script. The tool flags facts it is unsure of, and you can ask for a web check. You are still the editor.

Step 5: The publish pack

When a film is done, you also get what you need to upload it: title options, three thumbnails, a description with chapter timestamps, tags, a pinned comment and an SRT captions file, in one zip. The topic appears on the thumbnail as a short label.

Download the MP4 and upload it yourself. If you chose the realistic look, use YouTube's label for altered or synthetic content.

A finished example: a 5-minute documentary

One film made this way, "9/11 and Its Roots: 32 Years in 5 Minutes." Five minutes in four chapters:

  1. A Tuesday Morning in Boston (0:00)
  2. The Road to Al-Qaeda (1:04)
  3. 102 Minutes (2:30)
  4. 3,520 Days (3:59)

The public watch page for the film, with the 5-minute video, its chapters and a Make your own button

The public watch page. Anyone with the link can open it, no account needed.

Watch the film

Three thumbnails made for the film: "9/11 1979 to 2011", "Boston Control" and "102 minutes"

The thumbnails that came with it.

It is a heavy subject, and it shows the structure that works for long films: a human-scale opening, a clear chain of cause, and numbers placed where they matter.

Back to Noah Morris

One borrowed audio clip. One strike. A whole group of channels gone, close to $50,000 a month with them. His fix afterwards was to split every channel into its own company and its own AdSense account, so one strike could never take the rest.

That is the whole lesson of this pipeline, in three lines:

  • Your scripts are yours. Claude writes the brief, the tool writes the script from the brief, the originality score checks it against the source. Nothing is lifted.
  • Your pictures and voice are yours. Every shot is drawn for the line over it. The narrator is generated. Real people appear as licensed photos with a credit.
  • No two videos are the same. YouTube's policy calls repetitive, mass-produced videos "inauthentic content" (renamed July 2025). Thirty films that could be swapped for each other is the one way to lose everything above.

Thirty is a stretch goal. Three good videos beat thirty forgettable ones. But three good videos a week, for a year, is how every number at the top of this article was built.

Start

400 free credits and no card at getziztu.com/ai-youtube-video-maker. Ask Claude for ten ideas, pick one, paste its brief, and read the script it writes before you decide to make the film. Then tell me what it was about.

Top comments (0)