Getting the transcript
Reading the captions from YouTube. A video nobody has opened here before takes 10 to 30 seconds; this page fills in on its own.
Getting the transcript
Reading the captions from YouTube. A video nobody has opened here before takes 10 to 30 seconds; this page fills in on its own.

Luke Byrne (AI Luke) · @ai-luke
Words
2,461
Runtime
14:09
Speaking pace
174wpm
Reading time
10min
174 words per minute, between the 160 25th percentile and the 181 median of 349 measured videos. That distribution comes from the 349-video hook study.
Opening (first 30 seconds)
Believe it or not, these three videos are the exact same prompt. So, why do they look so different? Well, you gave it the prompt, right? And then you just hoped that it worked. Every time you're rolling the dice, and you're just going to use a bunch of credits. There's a much easier way now. Use Hex Field with Claude and Blender. Blender is an animation tool, but don't be scared, right? Because Claude is going to be using Blender for you. It's going to create
87 words, the words spoken in the first 30 seconds at 174 words per minute.
Free, no signup. See how the first 30 seconds hold attention, with rewrites.
Sentence shape
| Measure | This transcript |
|---|---|
| Sentences | 215 |
| Average words per sentence | 11.4 |
| Longest sentence | 40 words |
| Questions asked | 14 |
| Sentences containing a number | 17 |
Most used terms
Filler phrases
42 in total: actually 15 · like 11 · right? 9 · you know 3 · basically 2 · literally 1 · um 1.
A literal whole-word count of the same phrase list the Prepublish browser extension uses, so a phrase inside another word is not counted and a phrase used in its ordinary sense still is. It is a count and not a judgement.
Free, no account. See where attention is likely to drop, with a rewrite for each weak line. The free check shows the scores and the one issue costing the most. Or run it on the words above first.
Free · No login · See a sample audit first if you prefer.
What this transcript is
Every word below is the caption track YouTube publishes for this video, pulled from the video itself and reproduced unchanged. It is not Prepublish's writing, not a summary, and not a re-transcription: it is the video's own published captions. English captions, generated automatically by YouTube, in the video’s original language. Source: the video on YouTube. A channel that would rather this page did not exist can ask for its removal through the contact page, and it is removed.
No Script X-ray for this video: YouTube shows a Most replayed graph only once a video has enough views.
Believe it or not, these three videos are the exact same prompt. So, why do they look so different? Well, you gave it the prompt, right? And then you just hoped that it worked. Every time you're rolling the dice, and you're just going to use a bunch of credits. There's a much easier way now. Use Hex Field with Claude and Blender. Blender is an animation tool, but don't be scared, right? Because Claude is going to be using Blender for you.
It's going to create characters, environments, cameras, and it's going to turn gray blocks and its skins into motion. Taking blank files just like this one and turn them into the videos you saw at the very beginning. What if I told you you could get Claude to create the scene with the camera angles and everything in it that you want to happen? Get Claude to write the exact prompt and Hex Field to generate the video so that it's exactly what you want.
And you can just worry about the actual fighting scene and stop worrying about whether the camera is going to be in the right place or the hallway is the right size. And don't worry, you don't need to be a Blender expert to get this to work. For context, Blender I used for the first time ever like an hour ago. Okay, so let's get started. Let me show you how it's all going to work and then the downloads that you need.
So, we have here the scene builder, right? Then you're going to have the 3D animation layer on top of that scene. Then the animation layer on top, which then turns into image and then video, right? Boom. So, it's this whole layering sequence instead of this one-shot prompt, all starting from building the scene. So, the first thing you need to go and download is Blender. Link below for this, of course. Download Blender.
If you've never used it before, it's an animation software, but don't be scared off with this. You don't have to deal with any of it. That's why we're using Claude and Hex Field. Claude's going to handle this. It just gives an extra step so that Claude can actually try to build out this here before having to go and spend all your tokens and all your money on video generation. Speaking of which, head on over to the Hexedron.ai linked in the description below and you're going to come and download here this zip file that's going to then go into Blender so you can use Hexedron inside of Blender.
Now, once you've downloaded this here, you want to scroll right down till you find this bridge. This bridge essentially bridges you from the actual software Blender to use Hexedron. And for context, just a side note, you have this bridge for After Effects, Photoshop, Premiere Pro, DaVinci Resolve, Figma, and Blender. So, like everything, right? Anyway, scroll down here. We're going to copy that bridge, head on over to Claude's, and then we're going to go and add in our custom connector.
Claude, click on customize, connectors. Here you're going to click add and you're going to add in Hexedron bridge. Bridge. Bridge. And then paste this in here, click continue, and it will add it to your connectors. Once it's added to your connectors, you will then be prompted to sign in with your Hexedron account, and then it'll be all connected up. So, then the next thing you need to do is just open up Blender. Allow.
And then once you get Blender open, what you want to do is drag and drop in this Hexedron zip. Bam. And here, just click okay. Open this up. And now Hexedron will be here. See sign in to Hexedron. You can actually see it if you just press somewhere on here, press letter N, and you'll see Hexedron on the right hand side here. So, let's continue in browser and we'll go and sign in. So, click sign in, click allow, and now we can return to Blender.
And there we go. We are now connected up. So, if we look down here, what you can see is a scene builder, 3D model, character animation, image, video, camera, and assets. Now, all of these here are from Higgs Field. So, you can literally come here, describe a scene, use Opus 5 or Fable 5.1 now, and then let it generate this inside of Blender. For example, let's go to 3D model, and we'll say, "Generate a 3D car model." Click generate, and you'll see it'll pop up here on the right-hand side with all of the different details.
Again, don't worry about this just yet. Next, if we head on over to Claude Code, and then write a prompt something like this. Connect to Blender through Higgs Field Bridge, and build a 30-second pre-visualization in a single shot. A simple corridor of white walls, a box person walking down it, and a camera just following behind them. So, let's here make sure that we're on Ultra Code. Honestly, maybe even Fable 5 for this one.
Um but, I'll just leave it as Opus 5 for now. Right, so we have it now generated here. Actually, both things have generated at the same time. So, let me import this to scene. You can see that it's created this car. How amazing is that? I now have a car that's actually inside of this shot. And of course, you can then change it, etc. Turn that into image. You get the gist. Right, now if we go inside here of our room, whoa, there we go.
We've got our stick man, and he's inside of a room. And so, here you can be like, "Right, this hallway maybe feels a bit tight." So, you could then get it to expand it, construct it, make the guy taller or shorter, add more people. And all of this here is without you having to continually prompt an image generator. Instead, you're building the physics underneath it. And if you come here to Claude, you can see how he's actually doing this.
So, he's created this scene passport, and then he's setting this up, and he's actually using it right now. So, that's why See, it's automatically zoomed in. That was AI. That was Claude doing that, not me. Right, so now it's finished the whole walk path. We can press zero, take it to the camera, press space bar, and you can see that we then play the motion. So, this camera is from here. You can see the timeline along the bottom here.
You can click around, you can pause it, etc. Scroll it, scrub it, whatever it's called. And you can see then that motion, right? So, we have the initial motion. So, you could then replace this character with anything, and that should done this specific motion. But, speaking of changing characters, inside of Hex could you see Jin Jitsu, and Jin Jitsu as you get one motion, and you can continually change the character.
So, you can get like, for example, this guy, change it to this guy, etc., etc., right? We have an entire YouTube video on how exactly to do this. But, the cool thing with Blender is you can now create that motion without actually having to do it in real life. So, instead of having to go to a corridor, you can just do this. Right, so now let's make it a bit better than just a single shot, which by the way, of course, you can like preview inside of Cloud itself, which is amazing.
It's using Blender, and you can watch it happen. Anyway, so make sure to have this at the end of every prompt, and let me add in this bit here. So, we're going to say, "Now, make it more interesting. Add texture to the walls, bring the camera alive with different framings, and animate the lens. It moves continuously on its own route with accelerations and accents on the pauses. Everything smooth, no cuts, the hero always in frame." And then we'll send that in. >> [clears throat] >> And now it's going to go and modify this to a better version of this that we can actually work with.
Then after that, I'll show you how we can actually set this up so that we can have some textures, videos, and more stuff happening. Right, look at what we're working with so far. This is amazing. Look. Look at the texture on the walls and on the floor. And then look at this camera angle moving. All made using Cloud. How killer is that? It's still going on. Honestly, and the shadows and stuff. I'm so impressed. It's even the lights on the roof.
And And actually still working as well. So, this isn't even the final the final one. Right, so now we want to render this as a video. So, head on over here to output and then we've got 1920 by 1080. That's perfect. Frame rate's cool. Now, we want to change where it's going to go. So, I pretty much put everything in downloads, so we'll just go there. Click accept. Then we can give it a name. Actually, let's change the name to We'll call it one shot take.
Click accept. That's going to go there now. Media, video. So, then we want to go to encoding, change this video codec to MP4. Then we want to go to view, render play blast, and then this here will pop up. I'm going to head over to ChatGPT because we want some assets. Now, we want the person, we want the place, and we want the person that they're going to be fighting against. So, I'm just going to say, "Generate two people.
We'll do one ninja, one guy in suit, and a third image of a corridor in a hotel." Boom. Send that in. And so, here's the images. So, let's download them. Right, excuse the cat. She won't leave me alone. So, basically, add in these pictures we've generated now and that one shot take MKV, open that up, and now I'm going to say write a 30-second scene dance prompt based on the video blocking. Read the input video and write out the prompt second by second.
And we'll send that in. And then the output we'll put through Hex Field. Hopefully, no cat in that shot. Now, you want to grab this prompt, head on over to Hex Field, go to video scene dance 2.5, and then in here we're going to give it the assets and the prompts. Is that a good-looking guy? He's not me anyway. I'll tell you that much. Right, I've been kicked out of downstairs cuz the kids are asleep. Anyway, so back to this.
Basically, I went back here and realized I didn't even render the clip when it had finished. So, it wasn't even 30 seconds. It was 8 seconds. But, it's now 30. I've got a new prompt. I then went to chat GPT to turn my MKV into an MP4. And then we've grabbed that MP4, brought it here into Higgs food, right? So, let's see what it made. Looks good quality. Looks like the guy in the place. The camera's panning as we expected.
Ninja there. Oh, it's lights off. Hey, that's pretty cool. Look at this action. >> [laughter] >> That is so good. I actually can't believe the quality. That's nuts. And knowing exactly what happens now, all of the like angles make so much more sense. That was so good. Now, just just for context, let me check out the actual video itself. Yes, it slides up. Guy walking. So, yeah, it's just the different angles that are changing.
But, it's that wall angle's there. Yeah, and it's just added some extra details there. It's just used this as like the base inspiration. Oh, and something I want to check actually is let's see these side by side just to make sure the camera angle is what we want. >> [groaning] [groaning] >> And then the cool thing here is this video obviously has the motion, right? But, you can actually remove the motion. So, no 3D block.
And this is what it comes up with just with the exact same script, but no like default motion angle. So, already this is a different camera angle, which is pretty cool. But, it's completely Woah, completely different to what I was expecting. Holy moly. >> [groaning] >> And so, the thing here that you can then see is that sure this is pretty cool, but the angles are just completely different. And so, if you're trying to go for a consistent look and feel, maybe change characters, test some things out, you're just going to be burning through credits all day trying to get it to do this.
Now, for context when I'm adding this into Higgs fields here, you at the things you want to reference. So, we have these images uploaded, and then obviously I had the motion one uploaded to here, and you just go at, and then you reference the thing. So, this is video one, which is that video here. But, yeah, if you just hover over, you can see what it's talking about here. And you can save these as elements, so you can then consistently come back to them.
So, I could save this, you know, as hallway, and then now you can use this in different videos. This will be then the same consistent hallway, which obviously works great for character consistency. Behind say the blender here as well, as I was saying, you can, you know, do a bunch of stuff in here. You can generate the images you need, videos, scenes, models, etc. All inside of here, and then just get Claude to change what you want.
So, you never have to learn Blender. And honestly, the biggest hack here is that you never have to woo. You never have to now pay for another generation, another generation, another generation. It will just look the same every time cuz you know exactly the way the camera is going to go, the way the character is moving, and the scene behind it. So, there you have it. Make sure to check out Hexwood in the link in the description below.
Set it up with Blender and Claude so that you can then create consistent videos that you actually want to make, and not just some random video that you hope works.
The words are the caption track's own and nothing is reworded or re-transcribed. Paragraph breaks are placed between sentences so the text reads as prose.
Free tools for your own script: paste a draft and see where it stands before you record it.
Paste your draft and see where viewers are likely to drop off, with a rewrite for each weak line.
Paste the first 30 seconds of your own draft for a hook score and rewrites.
Check your draft against YouTube's advertiser-friendly guidelines before you record it.
Read this channel's public videos and transcripts, and download a writing brief for it.