How to Make AI Movies Like @mythrafilms Full Tutorial

Tech/Product/Workflow Summary (AI movie creation tutorial)

Channel/trend intro (social proof)

• Mentions a channel with ~50M views in 9 days and rapidly growing subscribers.
• The takeaway: this AI video format is performing well across many channels (not just one).

Core claim: AI movie making without hardware/software or appearing on camera

• The tutorial claims you can create these videos:
  • without a camera
  • without showing your face
  • without expensive software

End-to-end “master plan” to reduce manual work (15–20 minutes goal)

• Includes a ready prompt script (“master plan”) in the description.
• Designed so a large portion of the workflow happens automatically via prompts.

---

Step-by-step tutorial / tools mentioned

1) Get a master prompt + generate ideas (ChatGPT / Claude)

1. Download a prompt form containing blocks like:
   • “Copy Here” → “Copy Till Here”
2. Copy the prompt text into ChatGPT (“cloud”) and send it.
3. The AI generates 10 video topics.
4. Choose one topic by its topic number, or ask the AI to recommend the best topic.

2) Generate a consistent visual style

• After selecting the topic, the AI creates a visual style so the video maintains a consistent look.

3) Handle multi-part AI output

• If the AI stops mid-generation and asks “type next”, continue by sending “next” until it finishes.

4) Scene-by-scene story prop/prompt generation

• The AI writes video prompts/props per scene (cinematic elements for each situation).
• Scene 1 becomes the starting point for generation in an image-to-video pipeline.

5) Generate clips using “Google Flow” (Flow)

Instructions (as described):

1. Search Google Flow
2. Log in
3. Paste the scene prompt / image prompt from the AI
4. Set aspect ratio to 16:9
5. Select Omni Flash (as the generation option)
6. Click Generate

Notes emphasized:

• If Flow requests approval/permissions, read and approve.
• Validate each scene by generating it, then proceed to the next prompt.

6) Repeat per scene

• Copy the next scene prompt from Claude/AI.
• Paste into Flow and generate the next clip.
• Continue until all scene clips are generated.

---

Editing workflow (clips assembly)

7) Assemble clips in order using CapCut

• Open CapCut
• Create a new project
• Import all generated clips
• Arrange them on the timeline in order:
  • scene 1 → scene 2 → scene 3 → …
• Notes:
  • The tutorial suggests no separate voiceover/sound design is needed because the generated videos already include these elements.
  • Only minor trimming/cuts if necessary.
• Export the final video.

---

YouTube SEO / metadata packaging

8) Use a separate SEO/metadata prep script

• A “prep” script is provided in the description.
• Copy/paste the prep into the cloud.
• Send the full video script to Claude.
• Claude generates:
  • high-potential titles
  • descriptions
  • keywords
  • hashtags
  • other metadata

9) Thumbnail + packaging emphasis

• Strongly recommends preparing a good thumbnail and setting title/metadata correctly.
• Key viewpoint:
  • SEO alone won’t make a video viral
  • clickability depends on thumbnail/title
  • retention depends on story engagement

---

Main speakers / sources mentioned

• Tutorial creator (host; speaks throughout)
• Claude (AI for story/metadata steps)
• ChatGPT (alternative LLM option for topic/idea generation)
• Google Flow (cinematic clip generation per scene)
• CapCut (video editing/assembly)