**Этот плагин для Claude создаёт полноценные 2D-видео за минуты (Полное руководство) — Transcript & Summary | SozAI**
Source: https://sozai.app/transcript/claude-plugin-create-2d-videos/

Learn how to create full 2D animated videos quickly using a Claude plugin that automates script, animation, and video assembly.

## Key Takeaways

- Claude can fully automate 2D animated video creation from script to final MP4 output.
- The 'what if' style format is popular and effective for engaging viewers with animated storytelling.
- Voiceover quality and synchronization are essential for smooth animation and pacing.
- The plugin simplifies a previously fragmented process into one streamlined workflow.
- Video production is fast and scalable, enabling creators to produce multiple videos quickly.

## What the video covers

- The video showcases a Claude plugin that generates complete 2D animated videos from scripts and voiceovers within minutes to an hour.
- It explains that no image generators or neural network video tools are needed; Claude writes code to create characters, scenes, and animations.
- Examples of successful YouTube channels using this 'what if' style animation format are discussed, highlighting rapid view growth.
- The tutorial covers the entire process from topic selection, script writing, voiceover generation, to video creation using the plugin.
- Voiceover is crucial as the plugin syncs animation to speech; services like ElevenLabs are recommended for generating voiceovers.
- Instructions for installing and using the Claude plugin are provided, including setting project folders, choosing video format, resolution, and audio options.
- Claude automates tool installation (Whisper, FreeJS, FFmpeg, Node) and creates timestamps for syncing animation if not provided.
- Users can select visual styles and preview a short test clip before finalizing the full video production.
- The video creation process is efficient, producing a 6+ minute video in under an hour, allowing multiple videos to be made simultaneously.
- The plugin and prompts are shared via Telegram, and while a paid Claude subscription is required for full automation, parts of the workflow can be done with other tools.

## Chapters

1. 00:00 Introduction to 2D Animated Videos by Claude
2. 00:36 Examples of Successful 'What If' Style Channels
3. 01:10 Video Popularity and Viewer Engagement
4. 01:42 Challenges in Existing Tutorials and Solution Overview
5. 02:17 Demonstration of Claude Creating Animation and Sound
6. 03:15 Script Preparation and Topic Selection
7. 03:50 Voiceover Generation and Importance
8. 04:27 Installing and Setting Up the Claude Plugin
9. 04:59 Configuring Video Settings and Starting Creation
10. 06:05 Automated Tool Setup and Transcription

Answers

## Questions about this video

What does the Claude plugin do in video creation?

The Claude plugin automates the entire process of creating 2D animated videos by generating code for characters, scenes, animations, syncing them to voiceovers, adding sound effects and music, and assembling the final video file.

Do I need a paid Claude subscription to use this plugin?

Yes, a paid Claude subscription is required to run the full automated video creation process with the plugin, but individual steps like script generation or voiceover can be done with free tools.

How long does it take to create a 6-minute video using this method?

The process takes less than an hour to produce a 6-minute video, and multiple videos can be created simultaneously within the same timeframe.

## Full Transcript — Download SRT & Markdown

00:00

Speaker A

You've surely come across videos like these on YouTube or Reels. Simple hand-drawn characters, unusual situations, and animations that are hard to look away from. The most interesting thing is that you don't need image generators or neural network video tools to create these. The entire

00:14

Speaker A

footage is created by Claude. It writes the code that draws the characters, sets the scenes, and brings it all to life. This format is currently attracting a lot of attention. And if you look at how fast neural networks

00:24

Speaker A

are evolving, these videos will only get better. But while some save these clips thinking they should figure out how it's done, others are already launching channels and racking up millions of views. Take this English-language channel, for example.

00:36

Speaker A

There aren't many videos here yet, but the results are already impressive. Moreover, the authors didn't even invent the idea itself. They took a familiar "what if" format. And added new animation to it. You’ve surely come across such topics in Russian as

00:48

Speaker A

well. What if a person stopped feeling fear, or if gravity suddenly disappeared, and so on? The familiar theme looks different, fresh and new.

00:57

Speaker A

Instead of the usual montage of AI images, these are scenes where characters act, and what happens on screen helps tell the story. As of recording this lesson, the channel’s oldest video was released just a week ago and has already neared 3 million

01:10

Speaker A

views. Another one hit over 250,000 in less than a day. Of course, the script plays a role, but this delivery gives an extra reason to keep watching.

01:19

Speaker A

You're curious to see how the next situation will be shown. And this is not the only example. There’s also this channel. At the time of recording, it just appeared, released three videos, and also used a similar approach to

01:29

Speaker A

animation. At the same time, you don't have to make "what if" style videos. What if? Imagine short explanations of historical events, pop science, breakdowns of money, or how familiar things work. All this can be shown through clear, animated scenes. I

01:42

Speaker A

started looking for tutorials on how to create such videos. I found a few, but almost everywhere, there is the same problem. They show beautiful examples and talk about general things, but you have to piece the process together yourself. That's why I collected the

01:53

Speaker A

entire process into one Claude plugin. It handles the animation creation and video assembly. You choose a topic, prepare the script and voiceover, connect the Claude plugin, and start the process. Maybe you're thinking right now: "Okay, but how long will the

02:06

Speaker A

computer take to process and assemble it—will it take hours or a whole day?" Take a look at a clip we made in this lesson. The video itself lasts over 6 minutes. The computer made it in less than an hour, and I was completely free

02:17

Speaker A

the whole time. Drop tower falls 50. Everyone screams you yaw. The head teacher calls you on stage in front of the whole school. Everything you just saw, Claude created using code: the characters, the background, the movement. He also prepared sound

02:53

Speaker A

effects, background music, and assembled the result into a standard MP4 file. Next, I'll show you the entire process. I will post the plugin itself and the prompt for the script on Telegram. The link will be in the description under the video. Go ahead

03:03

Speaker A

and download the files. Once you have downloaded the files, open Claude. The process shown here requires a paid subscription. But even if you don't have one, the approach itself can still be useful. Individual steps can be tried using other tools. In this

03:15

Speaker A

tutorial, we will focus on video creation, so we will go through script preparation quickly. I have ready-made prompts for finding "What if" style topics and writing scripts. They will be on Telegram along with the plugin at the link in the description. But again,

03:28

Speaker A

you don't have to follow this specific theme or style. You can take the same approach and try it on your own channel and videos. First, copy the idea search prompt and send it to Claude. You can use the free version to generate topics

03:39

Speaker A

and text. We get a list of options. I will choose the topic: "What if you felt no fear at all?" Now, plug it into the script prompt. Send it and get the finished text. The foundation is ready.

03:50

Speaker A

Moving on to the voiceover. The voiceover is the main source material here. It is what the plugin will use to build scenes and sync animation to speech. You can use Voice Studio or another familiar voice generation service. In this example, I will use

04:03

Speaker A

ElevenLabs. Generate the voiceover, listen to the result, and download the file to your computer. At this stage, make sure you are satisfied with the voice, pronunciation, and tempo. From here on, the entire video will be built around this recording. If your service

04:16

Speaker A

can save text with timestamps, you can download that too. If such markers aren't available, the plugin will create them itself. Now, create a separate project folder and place the voiceover there. All the materials for the future video will be stored in this

04:27

Speaker A

folder. The next step is installing the plugin. In Claude, open the profile menu at the bottom of the sidebar, go to settings, and find the Plugins section. Next to the "Add" button, open the menu and select "Upload Plugin."

04:38

Speaker A

After that, upload the downloaded file. You can simply drag and drop it into the window. I already have the plugin installed. After installation, check that it has appeared in the plugin list. Now go to the code section and

04:48

Speaker A

connect the project folder where you put the audio. Next, select the model and the effort level. This example uses Opus. For effort levels, "High" or "Extra" work well, but I often find "Medium" sufficient. I don't enable the

04:59

Speaker A

maximum mode here; it consumes more tokens, and the chosen settings are enough for this project. Now, call the plugin and start the video creation. We need the full process from voiceover to the finished file. Add a simple request

05:11

Speaker A

. Create a 2D video based on the voiceover from the project folder. Send the request. Then the system will check which files are available. It found the audio I had and asked if there was a transcript. If you’ve prepared one in

05:21

Speaker A

advance, add the file to the folder or send the text to the chat. If there’s no transcript, ask it to create one with timestamps for every word. This makes it easier for the system to sync the action with the speech in the frame

05:32

Speaker A

. Next, choose the video format: horizontal or vertical. I usually make long videos for YouTube, so I choose horizontal. For a short vertical video, you would choose a different option.

05:40

Speaker A

Next is the resolution. In the process shown, options range from 720p to 4K. The higher the resolution, the longer the processing might take. So, for this demonstration, I’ll stick with 1080p.

05:51

Speaker A

All that’s left is to decide what to do with the audio. You can have the system prepare music and effects, leave only effects, or add your own music. I will choose automatic music and sound design generation. Now Claude will

06:05

Speaker A

prepare the tools needed for the task. In my case, it starts with Whisper to transcribe the voiceover and get the timestamps. For the visual part, it uses FreeJS, as well as FFmpeg, Node, and additional packages. In this process, Claude installs and configures

06:19

Speaker A

them itself. I didn’t need to go through each step manually. If there are overly long pauses in the recording, you can ask it to trim them separately. After preparation, the system will offer visual styles: DAC, Midnight, Golden, Neon, Storyboard.

06:32

Speaker A

Choose a suitable option or describe your own style if you already know how the video should look. I’ll go with one of the suggested styles. Next is an important point. First, Claude creates a short test clip of about 30 seconds.

06:45

Speaker A

This allows you to check the design, pacing, and sound before assembling the full video. Take a look at the result.

07:08

Speaker A

In my opinion, it turned out quite well. There’s already animation, music, and sound effects. If you aren’t happy with a specific scene, describe exactly w

07:20

Speaker A

the test clip satisfies you, ask Claude to continue and assemble the entire video. And meanwhile, our video is ready. The actual editing of this video took just over 5 minutes. Directing and planning the entire animation, along with Claude’s preparation, took about

07:34

Speaker A

half an hour. So, a video over 6 minutes long was ready in less than an hour. It’s important to understand that you can launch dozens of such videos simultaneously, and they will all be ready in an hour. Also, my setup

07:46

Speaker A

didn’t require any particularly powerful hardware. The result is a finished video with animation, music, and effects. It’s ready to be published. The entire process takes place within your Claude subscription.

07:57

Speaker A

Try assembling your own video and see what you get in your niche. Let me know in the comments which channel you would use this animation for. Thanks for watching. See you in the next video. M.

Topics: Claude AI 2D animation video creation animated videos AI video tools script to video voiceover synchronization YouTube animation what if videos AI plugin tutorial

Study this video

- [Flashcards from this video](https://sozai.app/tools/ai-flashcard-generator/?from=transcript&id=80710)
- [Quiz on this video](https://sozai.app/tools/quiz-generator/?from=transcript&id=80710)

Free, made by AI from this transcript. No signup.


---
This is the markdown twin of https://sozai.app/transcript/claude-plugin-create-2d-videos/ — the same content, without the markup.
Published by SozAI (https://sozai.app). Reuse and quotation are allowed with attribution and a link back.
Machine-readable index: https://sozai.app/llms.txt · data API: https://sozai.app/api/