GPT 6 Astra Is INSANE for Motion Design

CChase AI
Computing/SoftwareAdvertising/Marketing

Transcript

00:00:00GPT-6 Astra is the greatest model we have ever seen
00:00:03when it comes to motion design.
00:00:05And today, I'm gonna show you
00:00:06how to unlock its full potential.
00:00:08When you add this motion design skill to GPT-6,
00:00:11you can just go to Astra and give it a simple text prompt
00:00:14or a reference image, or even a link to a website,
00:00:17and it will create motion graphics for an explainer video,
00:00:20a product ad, or whatever else you want.
00:00:22So today, I'm gonna walk you through three examples.
00:00:25I'm gonna show you how you can set this up for yourself,
00:00:27and I'm gonna break down
00:00:28how this is all working under the hood,
00:00:30so you can customize it however you want.
00:00:32Now, what makes this all work is the motion design skill.
00:00:35Now, this video's all about how this works with GPT-6 Astra,
00:00:38but this can ultimately work with any AI model.
00:00:41So if you're using Cloud Code, this skill also works.
00:00:43And what happens is this gives your AI model
00:00:46a motion library, a bunch of different videos
00:00:49to essentially base its creation off of,
00:00:52so we never start from scratch.
00:00:54So for example, let's say I wanted to do a product video
00:00:57that shows sort of the inner workings
00:00:58of my particular product.
00:00:59Well, we have something that does exactly that.
00:01:01So if I click on the exploded product video,
00:01:05we get this.
00:01:15Now, chances are you aren't going to be making
00:01:18an exploded product video for a watch.
00:01:20However, if you had something else
00:01:22where you wanted to expose essentially the inner workings
00:01:24of some sort of product,
00:01:25well, we would use this as a template
00:01:27and you wouldn't have to come up with the prompt itself.
00:01:30That's already pretty much made for you with the skill.
00:01:33It then uses this as the template
00:01:35for your particular creation.
00:01:37And that idea is extended across
00:01:38all of these different styles.
00:01:40And if you want to create your own sort of style,
00:01:42it can, again, use the prompts
00:01:44that are sort of behind each of these
00:01:46as a template as well.
00:01:47So the idea is we never work from scratch.
00:01:49We're building off the things that we know actually work.
00:01:51Now, the other half of the equation
00:01:52is the Higgs field MCP
00:01:54because the skill is going to have the reference video.
00:01:56It's going to create the prompts for you
00:01:57and then it's going to send it off to Higgs field,
00:01:58which is going to use CDance 2.5
00:02:00to actually create the motion graphics.
00:02:02If we are inside of Astra,
00:02:03it has the ability to create images on its own.
00:02:05But if you're using this inside of Cloud Code
00:02:07or Fable 5.1,
00:02:08yeah, we're going to need to use Higgs field
00:02:10to create the reference images as well.
00:02:12Now, to do that, we're just going to go to Higgs field.ai,
00:02:15go to MCP and CLI.
00:02:16Either the MCP or CLI, both of these work.
00:02:19You're just going to click for your platform of choice.
00:02:22So if we're using ChatGPT,
00:02:23you're just going to do the add the Higgs field plugin
00:02:26and you're going to walk through the authentication process.
00:02:28So it's just going to ask for your login,
00:02:29your password, all that stuff.
00:02:30And the great thing about the MCP,
00:02:32instead of having to do all the work
00:02:33inside of the Higgs field UI
00:02:34and the balance between that and Codex,
00:02:37I can just sit inside of Codex
00:02:38or inside of Cloud Code
00:02:39and do everything from there.
00:02:40And the skill is super flexible.
00:02:42We use the SeedDance 2.5 as the default
00:02:44because that's the best AI video generation model out there.
00:02:46You can swap that for anyone you want.
00:02:48On the image side, it will default to GPT images too.
00:02:51But again, you can swap that to whatever you like.
00:02:54Furthermore, we can do text.
00:02:55We can give it reference images.
00:02:56You can even give it reference video.
00:02:58So really you just need to show up with an idea
00:03:00and a general idea of what you're trying to do.
00:03:03Then you need to decide what sort of style
00:03:05you want it to create it in.
00:03:06Again, you have these examples in front of you
00:03:08and then it does everything else.
00:03:09So it's pretty simple.
00:03:11You just got to show up with the idea.
00:03:13Now to set this all up is really easy.
00:03:14I give you the two prompts you can paste
00:03:16and decide of Cloud Code or Codex
00:03:17and then it's going to install the skill for you.
00:03:19I also added a set of guide specifically
00:03:21for Higgs field as well.
00:03:22Then all you have to do is hop into Codex,
00:03:24do forward slash motion design to call on that skill
00:03:27and give it your prompt.
00:03:29And here's the three demos I created using this skill.
00:03:32So the first one is a 2D explainer video talking about how noise cancellation works.
00:03:37It not only does the motion graphics, it also does audio.
00:03:40Now you can see down here, this is the prompt I used, right?
00:03:43It just said, I just said, use motion design to make a 15 second flat vector explainer about how
00:03:48noise cancelating headphones work.
00:03:50Adapt the flat vector sample, show the mechanism, and then show me the full prompt before generating.
00:03:56And from that basic prompt, this is then the prompt that was sent off to the Higgs field MCP.
00:04:02And you can see this is a lot more in depth.
00:04:04We went from a 2D explainer, use this one to this full prompt that breaks down everything
00:04:09from the color palette to the hero, to the type, to the motion, to a beat sheet,
00:04:13what we want to happen on each and every slide.
00:04:15So instead of you having to show up with this, you just need the skill and a simple input.
00:04:21Now, the other cool thing about these videos is most of these weren't necessarily
00:04:24one-shot what Codex was able to do, what Astra was able to do, is it generates the video the
00:04:28first time, it compares the first generation to what your essentially prompt was, what we're
00:04:32actually trying to achieve here.
00:04:34And if it thinks one particular section needs some work, it will then regenerate using CDANCE 2.5.
00:04:40And again, if you're worried about credits or tokens or anything like that, you can reduce that.
00:04:43But it essentially is a built-in system to make sure we don't just one-shot and pray.
00:04:47Like it will keep adjusting it to get what you want.
00:04:49So here's a look at the first one.
00:04:51- Noise cancellation fights sound with sound.
00:04:54Microphones detect the noise.
00:04:57A processor creates an opposing signal.
00:04:59The speaker plays it.
00:05:01The waves combine at your ear, reducing the noise, especially steady hums.
00:05:06So not bad for essentially a one-shot on our end.
00:05:11And it did two generations and created this.
00:05:13It's set for 15 seconds on this first prompt.
00:05:15It can go up to, I believe, 30 seconds with CDANCE 2.5.
00:05:18So again, you have a lot of wiggle room for what's actually generated here.
00:05:21And that was just a pure text prompt.
00:05:23For our second demo, we built off that exploding product thing.
00:05:29Because remember, we have sort of the style gallery, like I talked to you before,
00:05:34where you have all these different options to choose from.
00:05:38And so for this one, we used the deconstructed watch template.
00:05:41And this was the starting prompt I gave it.
00:05:44Use motion design to create a 20-second exploded product film for this keyboard using the two images.
00:05:49And then I included two reference images that Astra created itself.
00:05:53So we have the keyboard and the keys.
00:05:56And then the shot of the internals.
00:05:58And from there, it created this.
00:06:19Pretty solid work.
00:06:22Lastly, we created this campaign video for Canva.
00:06:26So the prompt here was, "Use motion design to make a 20-second kinetic promo based on
00:06:32Canva.com's visual suite."
00:06:34So I gave it the link to this webpage, which is a specific Canva product, and was like,
00:06:38"Hey, can we turn this into a video?"
00:06:41So I said, "Show one fictional coffee shop campaign becoming a social post, a story or presentation.
00:06:45Adapt the kinetic typography example and end with one idea, every format."
00:06:50And this was the full prompt it made for us.
00:06:53Again, you can see it breaks it down for every single scene, goes into a soundscape,
00:06:58all that sort of stuff.
00:06:59And here's what we got on essentially a one-shot.
00:07:15Again, I think that's a pretty solid output.
00:07:24There are some things we can tweak here.
00:07:26But I think the biggest takeaway from all three of these is that I gave it pretty basic prompts
00:07:32at the beginning of every single generation, right?
00:07:35The starting input is nothing serious.
00:07:37And we're getting like legit outputs.
00:07:39If you had a much stronger vision of what you're trying to create,
00:07:42and we're willing not to just like one-shot this, but to iterate, iterate, iterate,
00:07:46you can create some pretty solid motion graphics.
00:07:48Whether that's something that's 2D or some sort of product thing,
00:07:50or a campaign with like this Kinect typography,
00:07:53or really anything else within the style gallery.
00:07:56Like you can get solid outputs with Astra.
00:08:00So is motion graphics completely solved with something like GPT-6
00:08:03and the motion design skill and the HIC2 and the MCB?
00:08:06No, no, it's not completely solved.
00:08:08There's still a lot of work that can be done.
00:08:09But if you compare where we're at now with where we were at even a few months ago,
00:08:13you can see the direction this is heading.
00:08:15And you can see the sort of floor raiser skills and models like this are.
00:08:19For those of us who are not motion designers,
00:08:21you can create some really solid outputs using Astra, Fable 5.1,
00:08:27and a skill like motion design.
00:08:29So make sure to check out down below to get the skill for yourself.
00:08:33Let me know how this works for you.
00:08:34I'd love to see how you're able to adapt it as well,
00:08:36because you can customize this to your heart's content.
00:08:38And besides that, I'll see you around.

Key Takeaway

Integrating the motion design skill with GPT-6 Astra and CDance 2.5 enables users to generate complex 15- to 30-second motion graphics and product videos from basic text inputs.

Highlights

  • Integrating the motion design skill with GPT-6 Astra allows users to generate motion graphics for explainer videos and product ads using simple text prompts, reference images, or website links.

  • The motion design skill provides AI models with a motion library and reference videos to base creations on, eliminating the need to start from scratch.

  • The Higgs Field MCP and CDance 2.5 video generation model handle the creation of motion graphics based on detailed prompts automatically generated by the skill.

  • The system includes an automated self-correction mechanism that compares initial generations against the original prompt and regenerates sections when necessary.

  • Generated videos can run up to 30 seconds using CDance 2.5, accommodating various content formats like 2D explainers, product teardowns, and kinetic typography promos.

Timeline

Motion Design Capabilities and Workflow Mechanics

  • GPT-6 Astra creates motion graphics for explainer videos and product ads from simple prompts or links.
  • The motion design skill equips AI models with a pre-existing motion library to prevent starting from scratch.
  • Pre-made templates like exploded product views supply ready-to-use structures for custom projects.

Adding the motion design skill to AI models like GPT-6 Astra or Claude Code unlocks automated generation for explanatory and commercial video content. Instead of designing from zero, the system relies on existing style templates and reference motion graphics to shape the final output based on user inputs or website links.

Integration with Higgs Field MCP and CDance 2.5

  • The Higgs Field MCP connects AI coding environments to video generation tools.
  • CDance 2.5 serves as the default video generation model behind the motion graphics output.
  • Setup requires installing the skill via Codex or Claude Code and authenticating through Higgs Field.

The motion design skill works alongside the Higgs Field MCP to translate reference data and instructions into final video renders. Users install the skill in environments like Codex or Claude Code with brief prompt commands, defaulting to CDance 2.5 for video generation and GPT images for visual assets.

Demo Generations and Self-Correction System

  • A basic text prompt generates a detailed breakdown covering color palettes, slide content, and beat sheets.
  • The system evaluates initial video generations against the prompt and automatically refines weak sections.
  • Demonstrations include a 15-second noise cancellation explainer, an exploded keyboard product film, and a Canva kinetic typography promo.

Simple text inputs expand into comprehensive production prompts specifying color schemes, heroes, and slide-by-slide actions. An automated review mechanism tests the first generation against target goals and initiates secondary generations using CDance 2.5 to correct errors before final delivery.

Community Posts

No posts yet. Be the first to write about this video!

Write about this video