Summary of "How To Create Long Ai Cartoon Movies For FREE"

Overview (goal of the tutorial)

The video explains how to create long-form (30+ minute) 3D-style AI cartoon movies for free by using a workflow that:

Generates a script/chapter plan with an AI prompt (Google Gemini)
Produces images per scene with an image generator (called “Google Flow”)
Turns each image into an animated talking video with a free alternative (called Kwain AI / “CoinChat/Coin” in the subtitles)
Assembles everything with CapCut (editing, transitions, and export)

Open a new tab and search Google Gemini.
Paste a long pre-written prompt into Gemini.
Fill in:
- Title of the movie/story
- Number of chapters (the chosen chapter count determines video length; examples mentioned: 30/40/50; demo uses 25)
- Setting (example: Nigerian village)
- Nationality (example: Nigerian)
- Accent (example: Nigerian accent; selectable)
After submission, Gemini outputs:
- Character descriptions/personas (e.g., mother, village elder, twins)
- A set of image prompts
- A set of animation prompts
- Speech lines for each character (under something like a “speech and animation block”)

Key output concept: Gemini structures each scene so it includes:

For each chapter/scene:

Copy the image prompt from Gemini.
Open “Lab Google Flow” (subtitles: “Lab Google Flow”).
Create an account if required.
Configure:
- Aspect ratio
  - 16:9 for YouTube long-form
  - 9:16 for TikTok/Instagram
- Number of images per generation: set to 2
Download the generated image (tutorial mentions downloading at “1K” quality).

These downloaded images then become input for animation.

Because the referenced animation tool (“Google Flow” animation) may require payment, the tutorial uses a free alternative:

Go to Kwain AI (also referenced confusingly as “CoinChat/Coin” in subtitles) and create an account (account link is sent by email).
Generate video using:
- 16:9 selection
- Create video → upload attachment (upload the generated image)
- Paste Gemini’s speech + animation block content
The platform generates the video and downloads it.
Import the animated clip into CapCut.

The tutorial highlights an inconsistency:

To create a continuous story:

When generating scene 2+, use the previous generated/animated image as a reference.
In the animation/image tool, add that reference image into the prompt (subtitles mention “add to prompt”).
This helps the next scene continue from the prior one (e.g., same character appearance/positions).

To avoid “rushing” and unnatural AI behavior, the tutorial uses these strategies:

Prompts are optimized so each sentence is ~10 words max.
The animator limits clip length (subtitles claim videos over ~5 seconds aren’t allowed in that system).
For dialogue between characters:
- Each line generates a new image prompt, but keeps the same scene composition/character positions.
- The speaking character alternates per line (person two replies after person one’s previous line).

If a generated scene/dialogue isn’t correct on the first try:
- regenerate until it matches
- keep importing the best versions into CapCut

The workflow is presented as a way to produce long AI cartoon movies by combining:

Speaker: “Steve” (referred to in subtitles as “Well, Steve…”; the host delivering the steps)

Google Gemini (script/scene prompt generation)
Google Flow / “Lab Google Flow” (image generation)
Kwain AI (free video/image-to-video animation alternative; also referred to as “CoinChat/Coin” in subtitles)
CapCut (video editing, transitions, export)