I Copied This Viral AI Farming Channel Making $12k/mo (Step by Step)

An AI farming channel only two months old made over $12,000 and 4M+ views in 28 days. It uploads daily, and videos run 17+ minutes, stacking ads. Uploads hit 1.9M, 600K and 600K views in a month. Only one other channel (Mexico Nippon) does the same thing.

What the niche really is

It’s not normal farming; it’s farming in the Sahara. Pears, avocados, kiwis and watermelons forced to grow where they shouldn’t survive. A tomato in your backyard is nothing; in the desert it’s a story, and it scratches an itch people can’t look away from.

Thumbnail trigger: scale. People struggling to hold fruit 10× its normal size.

Branding and titles

With the prompts in the checklist: channel names in ChatGPT (I picked Future Farms), a logo prompt generated in Pippit (better than Gemini for this), then a list of titles in the same crazy style.

Script: add a voiceover

The original channel uses no voiceover. Don’t copy that part. For a long-term, monetized channel, add narration.

Use my master script prompt (set the length and your title). It builds in a flash-forward hook: show the result in the first seconds (farmers holding giant cantaloupes, fields to the horizon), then go back to day one. That answers the doubt before it starts, and it’s a masterclass in retention.

The prompt writes one minute at a time on purpose, because full 20-minute scripts get vague. Type chapter two, chapter three… Each chapter has hook visuals (no narration) at the top and story visuals in 6-second chunks.

A unique voice

In ElevenLabs → Voices → + → Voice design, paste my voice-design prompt (loudness 50%, guidance 60%), and pick a sample. Save it (English, American, male, middle-aged). In text-to-speech: model v2, speed 1, stability 40–50%, similarity 70–80%, speaker boost on. Keep chunks under 600 characters.

Visuals with Veo 3.1

Kling 3.0 on Higgsfield or Google Flow both work. I used Flow: text-to-video, Veo 3.1 Fast, 16:9, one output. Pick the most scroll-stopping hook prompt first, then generate each story scene. Download and name them in order (scene 1, scene 2…).

Pro edit in CapCut

In CapCut:

1. Voiceover only first, and trim the silences.
2. Put the hook clip first, push the narration to start after it.
3. Sync scenes to the narration.
4. Remove watermarks in 3 seconds: zoom one clip until the logo disappears → Copy attributes → select all → Paste attributes.
5. Lower the clips’ own audio so SFX don’t drown the voice.
6. Black fade transitions at about 1 s → Apply to all.
7. Captions template, and calm background music, low.

Thumbnail

Paste your title into my thumbnail super prompt in ChatGPT, then drop the result into Gemini or any image generator. You get the same oversized-fruit look as the viral ones.

📝 Your notes on this video

Saved on this device only

Copy notesDownload .txt