HƯỚNG DẪN LÀM KÊNH STORY KỂ TRUYỆN KIẾM TIỀN VIEW NGOẠI KHÔNG GIỚI HẠN KÝ TỰ MIỄN PHÍ

Summary

The presenter demonstrates a low-cost workflow for producing long, AI-assisted story videos aimed at international YouTube audiences. The proposed business model is to create long-form content that can attract substantial watch time and, according to the presenter, may earn more from views in Europe and the United States than from Southeast Asia. No actual revenue, RPM, or profitability figures are provided.

Suggested Production Playbook

• Choose a story niche and target audience: Research which countries best fit the channel’s content and audience. The presenter favors English, Thai, or Japanese audiences and says European and U.S. views may have higher ad revenue.
• Research successful videos: Find stories with strong view counts and examine their structure. The presenter prefers examples with hundreds of thousands of views, though one cited example had about 5,600 listens.
• Develop an original script prompt: Use an AI model to analyze a successful story’s structure and generate a detailed prompt. Refine it with strong hooks, emotional stakes, suspense, and chapter-by-chapter progression. The presenter’s example uses romance and drama tropes, such as an ordinary female lead and a powerful male lead.
• Generate and review the story: Request five possible story concepts, select one, then generate chapters sequentially. The presenter recommends reviewing and improving prompts rather than accepting the first output.
• Create narration: Copy each finished chapter into a text-to-speech tool, preserving the script exactly, and select a voice that fits the story. The presenter stresses that story quality matters more than voice choice.
• Generate visuals: Create character descriptions first, then use them consistently in image prompts. The presenter suggests roughly one image per minute of video, while also noting that a smaller set of 10–20 images may be sufficient.
• Edit in CapCut: Combine the narration and images, set images to about 60 seconds each, and apply simple zoom or animation effects. Create a thumbnail using the story script as input.

Metrics, Targets, and Operating Guidance

• Typical video length: 30 minutes to 3 hours are presented as possibilities; the recommended format is approximately 45–60 minutes.
• Script length: 80,000–100,000 characters, divided into 15 chapters.
• Production estimate: The presenter says 15 chapters may produce roughly 45 minutes to an hour; a sample chapter’s narration took about 3 minutes.
• Visuals: Roughly 45–60 images for a 45–60-minute video at one image per minute. The presenter also suggests generating about 60 image prompts, or using fewer images if adequate.
• Titles: Aim for roughly 95–100 characters, with a curiosity-provoking hook.
• Research signal: The presenter recommends looking for story examples with hundreds of thousands of views. One video discussed had around 5,600 listens.
• Costs: The presenter advises avoiding paid tools where possible, citing potential charges from image-generation services. They mention a roughly $300 credit amount appearing in one tool but say they stopped using it. No full production budget or monetization results are shared.

Practical Takeaways and Caveats

• Build a repeatable workflow: research → prompt design → chapter generation → voiceover → visuals → editing → thumbnail.
• Keep character descriptions consistent across image prompts, and inspect generated spreadsheets or prompts for errors before production.
• Use free or low-cost tools where practical, and check billing and API access before connecting a payment card.
• The tutorial demonstrates analyzing and adapting high-performing stories. Copying another creator’s script directly could create originality, copyright, or platform-policy risks; the safer business practice is to use references for structure and produce distinct scripts.
• The presenter’s revenue claims are directional rather than validated: the video gives no RPM, revenue, audience-retention, or monetization-approval data.

Presenters and Sources

Presenter: Nghiện Xây Kênh (channel named in the video metadata).

Tools and services mentioned: ChatGPT, Grok, Gemini, YouTube Summary browser extension, Audio TTS, Meta AI, Google AI Studio, and CapCut.