I Copied this $13K/Mo YouTube Channel Using Claude Code + Higgsfield

A channel only a month old, with nine videos, 36K+ subscribers and 2M+ views in 30 days. At a conservative $5 RPM that’s likely $13,000+ a month. Another, The Zen, uses almost the same format: one video at 7.6M views and about $4K a month from one or two uploads a week. This style can be replicated with AI, with no scriptwriting, design or long edits.

Setup

1. Install Claude Code (desktop app).
2. On Higgsfield (50% off) → MCP & CLI → CLI, paste the three lines into Terminal.
3. Claude Code → Customize → Connectors → + → Add custom connector, paste the MCP link, name it Higgsfield, connect and allow.

Niche and a proven topic

The free doc lists the top 15 faceless explainer niches working right now. Find a winning channel in yours, sort by Popular, and pick a proven topic. I used What did ancient humans do all day before jobs existed? (2.7M views in a month).

Script in 2 minutes

Paste the script prompt into Claude with your topic and the viral video’s transcript (YouTube → Show transcript). Claude analyzes the hook, structure, pacing and retention mechanics, then rewrites it in completely new words.

Why timestamps matter

Paste the Claude Code prompt with your script, and it may complain there are no timestamps. Without them it treats each paragraph as one scene, giving about 30 images for a whole video. Fast-changing visuals are a big reason these channels hold viewers, so fix it:

1. Voiceover first. In ElevenLabs, don’t use default voices thousands of channels use. Go to Voices → New voice and design one or clone yours.
2. Transcribe it for free with Notta (120 min/month) or TurboScribe (3 files/day) to get exact timestamps per line.
3. Paste the timestamped script into Claude Code. Higgsfield generates an image per line (100+ for a full video).

Auto-named images

Please download these images locally and name each file after its timestamp (e.g. 00-00, 00-10).

Allow it, and you get a folder with every image named by when it appears.

10-minute edit

In CapCut, add the voiceover, then drag each image to the time in its file name (0 s → 6 s, 6 s → 13 s…). No listening back, no guessing. It syncs perfectly.

📝 Your notes on this video

Saved on this device only

Copy notesDownload .txt