Summary of "How to Create TRENDING Reels, Shorts & TikToks with AI (Step-by-Step)"
High-level summary
This video teaches a repeatable, AI-driven workflow for making trending short-form videos (Instagram Reels, TikTok, YouTube Shorts) in a “dark/psychology” style (example: a page called Zolon that grew to ~1.3M followers). The creator demonstrates how to:
- Research verified information.
- Convert research into punchy viral scripts using ChatGPT.
- Produce an emotionally convincing AI voiceover with Hume.ai.
- Source B‑roll via Pinterest guided by ChatGPT keywords.
- Assemble everything in CapCut with layered audio, filters, captions and finishing touches.
The presenter emphasizes using verified research (not made-up claims), treating AI as a productivity multiplier (not a magic bullet), and building consistent branding (voice, filters, caption style).
Step-by-step method
-
Pick niche and topic
- Niches that work well: psychology, philosophy, masculinity, wealth, body language, relationships, hidden history, subconscious mind, dark storytelling, etc.
- Example target: a psychology video about cheating / dark psychology inspired by the Zolon page.
-
Research verified information
- Use Google Scholar (scholar.google.com) to find peer-reviewed articles, university research and credible studies.
- Open relevant papers and copy the verified information for your script source.
- Do not invent or exaggerate claims—base assertions on the research you found.
-
Set up a ChatGPT project and constrain sources
- Create a new project in ChatGPT and paste the selected article(s).
- Use a prompt that instructs ChatGPT to only use the provided/verified article as its source (prompt templates are referenced by the creator in the video description).
- This produces factual content extracted from the research.
-
Gather top-performing competitor transcripts for style/context
- Download transcripts of 2–3 best-performing videos from the exemplar page (e.g., Zolon) to analyze tone, pacing, hooks and cutting points.
- Recommended transcript tool: Riverside.fm (or similar).
- Add those transcripts into the same ChatGPT project so the model can mimic cadence and structure.
-
Generate viral scripts with ChatGPT
- Combine the verified research and the high-performing transcript style so ChatGPT produces short, hook-driven scripts optimized for short-form platforms.
- Output should be tight, fast-paced lines with clear cut points for visual edits.
-
Create an emotional, natural-sounding voiceover with Hume.ai
- Use Hume.ai to generate a voice that conveys emotion (Hume can act the emotion behind the script).
- Choose a voice from the library (the presenter used a deep voice for this niche).
- Hume.ai offers ~10,000 free characters/month (as of the video) and allows generating and downloading the TTS file.
- Generate and save the voiceover audio for import.
-
Choose background music and SFX
- Pick music that matches tone (CapCut built-in music, Instagram audio, or other libraries).
- Add an attention-grabbing intro SFX (example: “scary intense riser”) to hook viewers immediately.
- Layer audio hooks with visual and text hooks.
-
Source B‑roll / clips
- Use Pinterest (filter set to Video) to find video clips and imagery—preferred for its variety of creative B-roll.
- Use a final ChatGPT prompt that analyzes the script and returns targeted keywords for each cut/segment (keywords to paste into Pinterest).
- Example keyword: “dark surreal character art” (helped find unexpected clips).
- Download clips via a Pinterest downloader and import into your editor.
-
Edit in CapCut (assembly & finishing)
- Create a new project, import voiceover and B-roll.
- Apply CapCut’s “music enhancement” filter to improve audio; adjust levels.
- Use ChatGPT guidance to time cuts to the voiceover and to choose clip types.
- Add a color/filter (adjustment) layer to create a consistent, unique look.
- Add vignette(s)—the presenter used three for a darker aesthetic.
- Layer SFX and music under the voiceover so the speech remains dominant.
-
Add captions and text design - Use CapCut’s auto-caption feature (paid, but recommended) or alternative captioning tools. - Choose a consistent font (presenter used “Thrive”), text shadow, and small animations (fade-in ~0.2s). - Edit captions line-by-line if necessary for timing and emphasis.
-
Review, iterate, and repeat - Rinse and repeat the workflow across multiple videos. - Use ChatGPT repeatedly to generate keywords, script variations, and thumbnail ideas. - Expect that AI speeds workflow but human direction, testing, and iteration are required.
Key practical tips / technical details
- Always verify claims with academic papers; don’t fabricate psychology facts.
- Save and reuse prompts and settings (prompts available in the video description).
- Hume.ai: emotive TTS, ~10k free characters/month (good for human-like delivery).
- Riverside.fm: useful for downloading transcripts of competitor videos.
- Pinterest: excellent for surprising, high-quality B-roll; always set filter to Video.
- Use a Pinterest downloader to save clips (many options available online).
- Use consistent filter and caption styling as a branding element.
- Layer audio, visual, and caption hooks for maximum attention.
- The creator estimates AI handles about 70% of the work but you must oversee creative decisions.
Core lessons and cautions
- AI is a productivity tool, not an autopilot for guaranteed virality—strategy, testing, and editing judgment remain essential.
- Voice authenticity matters—avoid robotic TTS that breaks immersion.
- Credit creators where possible (ethical practice).
- Using verified research protects credibility and avoids misinformation.
- Iteration and consistency (voice + visual filter + caption style) build a recognizable brand.
Tools, platforms and sources mentioned
- Google Scholar (scholar.google.com) — research / academic papers
- ChatGPT — script generation, keyword prompts
- Riverside.fm — download transcripts of existing videos
- Hume.ai — emotional text‑to‑speech voiceover (sponsor)
- CapCut — editing, music enhancement, auto-captions, filters/effects
- Pinterest — B‑roll and video clips
- Pinterest downloader tools — to save video clips
- Instagram, TikTok, YouTube Shorts — target platforms
Speakers / sources featured
- Alex — main presenter / creator (signs off: “My name’s Alex. Let’s grow.”)
- Zolon — example high-growth page used as the model/style reference
- Hume.ai — sponsor / tool provider
- Riverside.fm — tool used for transcript downloads
- Google Scholar — research source
- ChatGPT — AI assistant used for scripting and keywords
- CapCut — editing and caption tool
- Pinterest — B‑roll source
Note: the transcript contained inconsistent spellings of the example page name (Zolon / Zultium / Zium / Zion); the presenter’s intended example is the single high-performing page referenced.
Category
Educational
Share this summary
Is the summary off?
If you think the summary is inaccurate, you can reprocess it with the latest model.