I Pushed Claude Animations to the Max (Copy My Prompt)
A six-minute walkthrough of turning a single reusable prompt, a Pinterest reference folder and Higgsfield into coded UI animations, talking-head reels and AI sound design.
Posted
yesterday
Duration
Format
Tutorial
educational
Views
18.7K
647 likes
57 · 43
Big Idea
The argument in one line.
Great AI-made video comes from a saved, trial-and-error-refined prompt plus visual references, so you generate mock-ups first and only then let the model code and render the final animation.
Who This Is For
Read if. Skip if.
READ IF YOU ARE…
A video editor or freelancer who wants Claude to produce SaaS-style UI animations and talking-head reel edits instead of building every shot by hand.
A creator who keeps rewriting the same long prompt and wants a repeatable method for saving a style once it finally looks right.
Someone already using Claude with MCP plugins who wants to see Higgsfield wired in for image and footage generation.
SKIP IF…
You want a deep tutorial on Blender or DaVinci Resolve; this stays at the level of prompting and tool wiring.
You need the actual prompts reproduced here; they live in a Google Doc linked in the video description.
TL;DR
The full version, fast.
Zane Hoyer argues that polished AI video starts as one carefully refined paragraph-style prompt. He keeps a Google Doc of prompts for UI animations and talking-head reels, swaps in brand names, colors and images, and pastes them into Claude. For animations he gathers Pinterest references, connects Higgsfield through the plugins menu, has Claude generate six mock-up images, then approves them before Claude codes and renders the video. For reels he uploads footage and a style prompt. When a result is wrong he corrects it in chat, then asks the AI to write a new prompt from the conversation. Sound design runs through GPT 6.2 Astra in Higgsfield.
Free for members
Chat with this breakdown — free.
Sign in and you get 23 free chat messages on us — ask for the hook, quote a framework, find the exact transcript moment, generate a markdown action plan. Bring your own key when you want unlimited.
“You can think of this process kind of like the design team where we get everything how we want it to look.”
clear metaphor for the mock-up stage→ newsletter pull-quote↗ Tweet quote
04:36
“Once you get the outcome that you like and you want to save it for next time, you can ask the AI to create a prompt for everything you said in that chat.”
See every word as it's spoken — crank it to 2× and still catch all of it. The same dual-channel trick behind Amazon's Kindle + Audible.
17px
Every one of these clips you see started as a single paragraph. And in this video, you get access to everything. The exact prompts, the setup, and a few of the rules I follow that have people asking, how did he make that?
I spend insane amounts of time talking to AI. In fact, the last five of my videos have all been edited by Claude. This is coming from someone who edits for some of the biggest channels on YouTube.
Now, once you copy these prompts and change a few of the words, they become yours. Which in itself raises another question. If anyone can make something this good...
Is it still good? Now before we start, I'm going to drink some caffeine, I'm going to put in some nicotine, and I'm going to start playing this type of music. Now that we're in flow state, we can think clearly.
So if you go down in the description, you will see a link that will bring you to this document. And as you can see over here, there's multiple different tabs on this document for different types of videos we can make. These top ones here are for UI animations or software animations like you saw in the intro.
And these bottom ones are different styles for the talking head Instagram reels that you also saw in the intro. And by the way, all of these are designed so that you can change out the product names and the colors and the images very easily. And then down here, we have a sound design prompt.
We're going to be using this later with GPT 6 .2 Astra. It just came out and I used it for sound designing all of my clips and it is really good. now if you ever want to make your own style i'll show you how to do that as well essentially what these prompts are is the result of me going through a bunch of trial and error and correcting little tiny things multiple times every prompt and then for all the videos i get that look amazing i'll throw the prompt into here so i have clawed open here and i want to generate my source footage as well as really crisp clear images of the pre -rendered shots basically just so i can confirm it before it renders it out and so i'm just going to connect higgs field to my cloud really quick i've been using them a bit more lately because they've been integrating claude astra blender and davinci resolve into their ai workflows which are all things that i use i do have an affiliate link in the description so in cloud we just have to come to plugins we just have to click connect and then we're good to go the higgs field mcp lets us get a lot cleaner and more advanced starting images this is a before and after for the
difference. Some of these prompts require references, so we can just go to Pinterest. I'm going to search 3D SaaS videos, and then we can just essentially click through different style references until we get some that we like.
I think something like this would be cool if we built out a monitor. This one looks pretty cool, so I'll save that. This one's pretty good, I'll save that.
This one also looks pretty good. I'm going to create a new folder. I'll just title it something like 3D Reference.
And then I'll drag all of these into that folder. So now on the prompt page, I can go through these inputs and change the name of anything I need to change. So for example, the brand could be Rain Zane.
And I'm just going to go one by one and fill all of this out. And then once I do that, I can copy everything like this to my cloud, paste in my prompt, and it's just going to go like that. So we have our prompt, we have our references, and now I'm just going to tie everything together by saying, see my prompt above for the motion animation style i've also linked a bunch of references in my mac folder labeled 3d references i want you to use higgs field and create six mock -up images so that i can see what they're going to look like before we begin coding everything and rendering everything out you can think of this process kind of like the design team where we get everything how we want it to look and then once we confirm everything this prompt basically tells it to dissect the images and basically turn them into real coded video and so now while we wait for this we can open up another tab here And this one can be for our talking head reel.
You can either upload yours or your client's footage directly into this. Or for me, I don't have any footage right now. So I'm just going to generate something really quick.
Using Higgs field, can you generate me some footage of a man with a minimalistic setup? Talking to the camera, full body shot. I'm going to say use C -Dance 2 .5 4K.
And we can just wait for that. So here is what it came out looking like. So for the talking head reels, we can just upload this into here.
I can come to my reel styles. So I'm just going to select this one. And then again, copy all of this.
and then hit generate now i'm also updating this prompt so that it connects to your sound design library if you want it to so if we go back to our other video now here is the storyboard it laid out for us i'm just going to say okay let's turn it into a video now and the prompt we already gave it basically directs it on which motion to use so here's what it ended up giving us but i'll just show these at the end of the video now really quick if you're wondering how i get my prompts i'm just going to show you how i prompt it so that essentially going to show you the prompts on how i make my prompts As you can see, I've uploaded this raw footage as well as some assets for how I want my editing style to look.
I want you to dissect the raw footage, build out a transcript of what they say, and then I want you to create a storyboard of multiple animations on screen. And then I want you to create a storyboard for multiple animations and on -screen animations for when she's talking. And then from there, obviously, it's going to make some mistakes and it's not going to look perfect.
You just have to correct it over and over again and basically just tell it little things that you don't like. And then eventually, once you get the outcome that you like and you want to save it for next time, you can ask the AI to create a prompt for everything you said in that chat. So now that GPT 6 .2 Astra came out, you can either connect ChatGPT to DaVinci Resolve like I showed in my last video, and this will cut up your footage and sound design.
Or if we come to Higgsfield's supercomputer, we can just select it here. We can upload our video. I can upload my sound effects pack just like that.
And then we can come to this Google Doc that's in my description. And I can select the sound design prompt. And I'm going to copy this.
and paste it into here and then we can hit generate gpt6 astra was good for sound design before but now that we have 6 .2 astra it takes sound design to a new level now i'm going to continue to update this prompt list so it's just going to get better and better over time
The Hook
The bait, then the rug-pull.
Every clip in the intro started as one paragraph of prompt. Zane Hoyer, a professional editor, shows how he stores those prompts, feeds them to Claude with references, and corrects the output until it is good enough to save.
Frameworks
Named ideas worth stealing.
02:55model
Mock-up first, code second
Prompt plus references
Six mock-up images
Approve the look
Code and render
Generate stills to lock the design, then let the model dissect the approved images and turn them into real coded video.
Steal forany generated motion-graphics project
04:25model
Correct, then capture the prompt
Run the prompt
Fix small issues in chat
Repeat
Ask the AI to write the prompt
Iterate in conversation until the output is right, then ask the AI to summarize the chat into a reusable prompt.
Steal forbuilding a personal prompt library
CTA Breakdown
How they asked for the click.
VERBAL ASK
01:02link
“if you go down in the description, you will see a link that will bring you to this document”
The CTA is the prompt document in the description, and the Higgsfield affiliate link is disclosed on camera.
Add Modern Creator as a preferred source and Google shows you more of our breakdowns in Search, Top Stories, and AI Overviews. It only changes what you see, and you can undo it in your Google settings anytime.
Add to Preferred SourcesOpens your Google source preferences with us pre-loaded. Tick the box and you're done.
A five-minute walkthrough of how to get Claude to build a continuous-motion documentary entirely in code, starting from a Pinterest reference board and a few rounds of blunt feedback.
A freelance editor shows the folder structure, three AI agents, and Instagram offer he says lets him run a six-figure editing business in one to two hours a day.
A screen-recorded walkthrough of Higgsfield's Vibe Motion tool, where a creator turns plain-English prompts into a dozen client-ready motion graphics in minutes.
A five-year YouTube editor names the four rules behind his channel's intros, then rebuilds one, badge, thumbnails, cursor, and glow, live inside DaVinci Resolve's Fusion page.