AI YouTube Automation

Open full screen · recording mode — sized for vertical video.

1. 01

   Pick a subject you can keep up

   Free sample

   One subject, one format, and your first ten titles.

   In this step you choose something you can still stand in six months. The failure mode in this strategy is not quality, it is quitting: channels die at video nine. Picking a subject where you already have curiosity, and a format you can produce in a repeatable way, is what gets you past that.

   1 — Choose subject and formatCopy prompt

   I want to build a faceless YouTube channel with AI-assisted production.
   
   Subjects I can stay interested in for a year: [YOUR SUBJECTS]
   Time per week: [TIME]. What I can do: [YOUR SKILLS]
   
   Per subject:
   1. Which faceless formats fit — narration over stock, data explainer, list, documentary-style, tutorial.
   2. What the viewer is actually doing while watching, and what that means for pacing.
   3. Whether the subject has enough material for fifty videos, or runs out at twelve.
   4. What makes a channel in this subject look interchangeable with every other one.
   5. What the barrier to entry is — and if there is none, why anyone would watch mine.
   
   Then pick the strongest combination of subject and format for my time budget, and write the first ten video titles.

   Claude. Ask it to argue against your favourite subject — the objection is usually the real constraint.

   [YOUR SUBJECTS]

   Real interests. Fifty videos about something you do not care about is not happening.

   [TIME]

   Honest weekly hours. One video a week at four hours beats three a week for a fortnight.

   Then: Write the first ten titles into a list you keep. Never start a video without the next one already titled.

2. 02

   Set up your channel

   In the package

   Name, art, about and the rules you will stick to.

   In this step you build the shell. It takes an afternoon and it is the difference between a channel that looks like a project and one that looks abandoned at birth. You also write down the platform rules that apply to AI-assisted content now, so you are not rebuilding later.

3. 03

   Write your first script

   In the package

   A script with sources, an opinion, and a structure that holds.

   In this step you write the video. Research first, from sources you choose, because the part that makes a channel worth subscribing to is a perspective — and a model with no sources produces the same summary as everyone else's channel.

4. 04

   Make the title and thumbnail

   In the package

   A title and thumbnail that make one promise, checked at small size.

   In this step you make the part that decides whether the script is ever heard. Title and thumbnail are one unit: the thumbnail shows the situation, the title supplies what it cannot say. The test is whether both still work at the size of a phone in a list.

5. 05

   Publish your first three videos

   In the package

   A pipeline you can repeat, and three videos live.

   In this step you turn one video into a process. Three videos is the minimum that tells you anything and the point at which the pipeline either holds or collapses. You end with a chain you can run weekly without re-deciding everything each time.

6. 06

   Read the numbers, plan ten more

   In the package

   Where viewers drop off, and ten titles queued with the fix.

   In this step you use the only feedback that matters early: the retention graph. Views tell you about the thumbnail, retention tells you about the video. You end with a specific change and ten more titles, which is what keeps the channel alive past the point where most stop.

What you need

core means you cannot skip it. The rest only if you need it.

• Descriptcorefree planEdit video by editing text
• ElevenLabscorefree planVoice, including a clone of your own
• NotebookLMfree planResearch grounded in sources you choose
• Perplexityfree planFast sourced research with citations
• Canvacorefree planThumbnails without a designer
• Opus Clipfree planShorts cut from long-form
• Vizardfree planClip selection and repurposing