FREE Claude Prompt That Creates VOX Style Videos AUTOMATICALLY | Here's How — Transcript
Full transcript
- 0:00Now, there's this crazy new trend going
- 0:01viral, like crazy, that features Vox
- 0:04style animations. Those paper cut style
- 0:07visuals used to tell documentary stories
- 0:10and keep viewers hooked and glued to the
- 0:11screen. And they're pulling in some
- 0:14crazy numbers. So, I checked YouTube to
- 0:16see if anyone had covered how to make
- 0:17these with free tools. But all I could
- 0:20find were complicated tutorials using
- 0:22Claude code or expensive paid software.
- 0:25Nothing beginner-friendly at all. So, I
- 0:27did my own research and I found a way to
- 0:29create these animations completely free
- 0:31without paying for any software. So, in
- 0:33this course, I'm going to walk you
- 0:35through exactly how to create
- 0:36high-quality Vox style animations from
- 0:38scratch using 100% free AI tools. And
- 0:42simple enough that any beginner can
- 0:44follow along. I mean, check out this
- 0:45cool animation I made in like 5 minutes
- 0:47with this free strategy.
- 0:49>> May, 1925.
- 0:52Paris.
- 0:53A well-dressed man reads a newspaper
- 0:55story about the Eiffel Tower rusting and
- 0:58costing a fortune to maintain. His name
- 1:01is Victor Lustig.
- 1:03He has an idea.
- 1:05Days later, forged government stationery
- 1:08in hand, he invites five scrap metal
- 1:10dealers to a private suite at the Hotel
- 1:13de Crillon.
- 1:14He calls himself a deputy director of
- 1:17the Ministry of Posts. The city, he
- 1:20explains, can no longer afford the
- 1:22tower. It will be sold for scrap.
- 1:25Quietly, to avoid public outcry.
- 1:28He picks the most eager dealer, a man
- 1:30named Andre Poisson.
- 1:32He takes the payment and a bribe on top
- 1:34of it.
- 1:35Then he boards a train to Vienna.
- 1:38Poisson,
- 1:38>> [music]
- 1:39>> too ashamed to admit he was fooled,
- 1:41never goes to the police. So, 1 month
- 1:44later, Lustig returns to Paris.
- 1:47He gathers a new group of dealers. He
- 1:50sells the tower again.
- 1:51>> All right. So, for this walk-through,
- 1:52the very first thing you're going to
- 1:54need is access to the master prompts
- 1:56we're going to use to automate this
- 1:57entire process. You should find that
- 1:59either in the description or in my
- 2:01completely free school community. Once
- 2:03you've got the master prompts, the next
- 2:05thing you need is a chatbot. Now, there
- 2:07are a couple of options here. For the
- 2:08completely free users, Deep Seek should
- 2:11be your go-to. Deep Seek is a
- 2:12Chinese-launched chatbot that doesn't
- 2:14have a paywall. It's just like GPT, but
- 2:16you don't have to pay for it. Another
- 2:18good option is my personal favorite,
- 2:20Claude. And the last one is the OG,
- 2:23ChatGPT. You can pick any of these to
- 2:25work with. They all excel in different
- 2:27places. For this walkthrough, I'm going
- 2:29to go with Claude, but you can still use
- 2:30Deep Seek or GPT. They both work the
- 2:33exact same way. So, I'm just going to
- 2:34choose Claude, head back to my master
- 2:36prompts, scroll down to the bottom, hit
- 2:39command plus A to highlight everything,
- 2:41then control plus C to copy it all. Once
- 2:45I've copied it, I'll head back to
- 2:46Claude, paste it in, and send it. Now,
- 2:49once you send it, Claude thinks for a
- 2:50bit and then asks you for the source
- 2:52PDF, which is basically the content
- 2:54engine we're going to use for this whole
- 2:56process. [music]
- 2:56To get the source PDF, just head back to
- 2:58the master prompts. I stored everything
- 3:00in there, and scroll down to the engine
- 3:02prompt. Click on that. Once you do, it
- 3:05takes you to an eight-page PDF on a
- 3:07Google Drive. Just click download to
- 3:09save it to your computer. Then, head
- 3:10back to Claude, drag it into the prompt
- 3:12box, and type engine attached. Once you
- 3:14send that, Claude proceeds to stage one.
- 3:17So, now Claude asks you what niche you'd
- 3:19like to make these animations in.
- 3:21There's history, there's money and
- 3:23power, there's disaster and survival,
- 3:26there's documentary, tech, and sports.
- 3:29But, from what we've seen from Vox,
- 3:30they're mostly make stuff around crime
- 3:32and documentaries. So, I'm going to
- 3:34choose number one, which is the default.
- 3:37But, you could pick whatever style you
- 3:38want. And here's the thing, even if the
- 3:40style you want isn't listed here, you
- 3:42can just type your own style and the
- 3:43model will still proceed with the next
- 3:45steps, which is creating video ideas for
- 3:48whatever niche you chose. So, I chose
- 3:49the the and documentary niche, and it
- 3:51gave me 10 ideas to pick [music] from
- 3:53for a specific crime-based video. I'm
- 3:56just going to go through them real
- 3:57quick. All right, for this one, I think
- 3:59I like idea number four. The man who
- 4:01sold the iPhone 12 twice.
- 4:04Sounds interesting. So, I'm going to
- 4:05copy that, paste it into my chatbot, and
- 4:08send it.
- 4:09Once I do that, the next thing Claude
- 4:11does is ask me how long I'd like my
- 4:13script to be. Now, you can choose
- 4:14anywhere from 30 seconds to 20 minutes.
- 4:17It just depends on the kind of video you
- 4:19want to make.
- 4:20For documentary videos, you should aim
- 4:22for around 10 to 20 minutes. That's the
- 4:24sweet spot for ad revenue. But, since
- 4:26this is just a brief walk-through, I'm
- 4:27going to go with the 1-minute video.
- 4:29Once you choose that, the chatbot takes
- 4:31in your request and gives you a precise
- 4:331-minute long script, along with the
- 4:35exact word count it created. Now, you've
- 4:37got your script. For the voiceover,
- 4:39there are platforms like Eleven Labs and
- 4:41a couple of other free options, but
- 4:43we're going to get into that later on.
- 4:45For now, let's just type next to move to
- 4:47the next step. One thing to keep in
- 4:49mind, depending on the chatbot you're
- 4:50using, this can be a little different.
- 4:53Claude is pretty strict. I used Deep
- 4:55Seek before and typing something like
- 4:57proceed would work fine for next. But,
- 5:00with Claude, you'd have to type voice or
- 5:03proceed for it to actually move on. Once
- 5:06it does, it starts converting the script
- 5:08it just gave you into visual beats that
- 5:10we'll use to create the images for your
- 5:12specific video. For our 1-minute script,
- 5:14we ended up with about 27 beats based on
- 5:16what Claude detected, which is actually
- 5:18pretty accurate since this involves a
- 5:19lot of movement. This looks pretty good.
- 5:22You can ask it to give you a specific
- 5:24number of beats, but I'm happy with
- 5:25this. So, what I'm going to do now is
- 5:27type next.
- 5:29When I do that, the model starts
- 5:31creating a very detailed TXT file that
- 5:34we're going to use to create our images
- 5:35automatically. All right, once you get
- 5:37your TXT file from Claude, the next
- 5:39thing you need to do is download it.
- 5:41This part is really important because
- 5:42we're going to use it to automate the
- 5:43image creation. Depending on how long
- 5:45your script is, you might need to
- 5:47generate up to 50, 100, maybe even 200
- 5:50images. With this TXT file, all you've
- 5:52got to do is head over to a platform
- 5:54called Google Flow. Click on create a
- 5:56new project. Once you're in, click on
- 5:58the video model at the bottom and make
- 6:00sure you switch from the video models
- 6:02over to the image [music] model. Google
- 6:03Flow gives you access to both.
- 6:06Once you've switched to the image model,
- 6:07select the aspect ratio you want. For
- 6:10this, I'm going with 16 by 9 because
- 6:12we're creating our images in the
- 6:13long-form aspect ratio. Then for the
- 6:15model, I'm going to choose Nano Banana.
- 6:18Now I need a way to batch produce images
- 6:20here automatically. For that, you're
- 6:22going to need an extension called Zappy
- 6:24Flow made by a creator on here. Really
- 6:26nice extension. Head over to the Chrome
- 6:28Web Store and click add to Chrome. Once
- 6:30it's downloaded, you should see a pop-up
- 6:32appear on the right side of your screen.
- 6:35Once you've got that pop-up, head over
- 6:36to Google Flow and make sure it shows
- 6:38connected to your Google Flow project.
- 6:39Down here, you can customize how you
- 6:41want the images to be created. The
- 6:42default settings work just fine. The
- 6:44only thing I'd say you should turn off
- 6:46is include serial number in file name.
- 6:48If that one's on, just turn it off. Now
- 6:50what you've got to do is upload the TXT
- 6:52file you got from Claude. Upload it
- 6:54right in here. According to Claude,
- 6:55we've got 27 beats, which means Claude
- 6:57made 27 image prompts for us. And as you
- 7:00can see, we've got 27 prompts here. What
- 7:02I'm going to do now is scroll down and
- 7:03click on start generating. And just like
- 7:06that, the Zappy Flow extension starts
- 7:08creating my images completely from
- 7:10scratch.
- 7:11I could literally just leave this here
- 7:12and go do something else while the
- 7:14extension automates the whole thing for
- 7:16me. And the cool thing about this
- 7:18extension is that once it creates your
- 7:20image, it saves it straight to your
- 7:22computer. So you've already got all your
- 7:24images generated and saved
- 7:26automatically, and you can just leave it
- 7:28running and go do something completely
- 7:30different. Pretty cool. All right, so
- 7:32after a couple of minutes, we've now got
- 7:34every single image we need for our
- 7:35video. As you can see, the extension
- 7:37generated all 27 images.
- 7:39>> [music]
- 7:40>> Not going to lie, this is really, really
- 7:42good. Everything's generated in the
- 7:43style we want and I'm genuinely blown
- 7:45away. One thing I like is that if we
- 7:47click on the downloads folder, we can
- 7:49see every single image that's been
- 7:51downloaded and saved to our computer,
- 7:53which is really cool. Now that we've got
- 7:54our images, we can close the extension.
- 7:57Next, we need to animate these images by
- 7:59turning them into videos. For this,
- 8:01we're going to head back to Claude and
- 8:02type next. Once you do, Claude starts
- 8:05creating a very detailed universal video
- 8:08generation prompt. Now, the default is
- 8:1010 seconds, but here's what I recommend.
- 8:13Ask Claude to make the duration less
- 8:15than 10 seconds because we're planning
- 8:17to use free software. For video
- 8:19generation, the more time an image has
- 8:21to animate, the more it costs. A
- 8:2310-second animation
- 8:25>> [music]
- 8:25>> is going to cost more than a 5-second
- 8:27one. So, what I like to do is head back
- 8:29to Claude or whatever chatbot you're
- 8:31using and ask, "Could you change the
- 8:34duration to 5 seconds instead of 10?"
- 8:37Then send it. When you do, Claude gives
- 8:39you a new prompt. Of course, you can
- 8:41still use the 10-second prompt if you're
- 8:43planning to use a paid software, [music]
- 8:44but for the free users, I think 5
- 8:46seconds works just fine. All you got to
- 8:48do is copy the entire video prompt. Once
- 8:50you've copied it, head over to Flow,
- 8:52click on the Nano Banana option, and
- 8:54switch from the image model to the video
- 8:56model. Then select one output. For the
- 8:59duration, select about 6 seconds. Once
- 9:01you do that, start dragging the images
- 9:03into the prompt interface. Paste in the
- 9:05prompt you got and send it. Flow then
- 9:07starts creating your video from scratch.
- 9:09Now, all you've got to do is repeat this
- 9:11exact same process. Drag in the image,
- 9:13paste in the prompt, send it, go back to
- 9:15your chatbot, grab the next image, drag
- 9:17it onto the prompt box, paste it in,
- 9:19send it, and [music] just keep doing
- 9:20this until you've dragged every single
- 9:22one of your images through so we can
- 9:24generate all our videos. Now, while Flow
- 9:26is handling this, I like to compare my
- 9:28results. The model we're using right now
- 9:30on Flow is the Omni Flash model. So,
- 9:32while it's generating the videos for us,
- 9:35I want to try out SeaArt Dance and see
- 9:36what it can do. I'm going to access
- 9:38SeaDance through Higgsfield.
- 9:40Full disclosure, [music] Higgsfield is a
- 9:42paid platform, so that's something to
- 9:43keep in the back of your mind. I'm going
- 9:44to click on the video panel option once
- 9:46I sign up. Then I'll click the upload
- 9:48button on Higgsfield and upload the
- 9:50images I want to animate. Now, ZappyFlow
- 9:52saves all the assets it generated in a
- 9:55folder. So, just head over to that
- 9:57folder and upload the image you plan to
- 9:59animate. Once you do that, select the
- 10:00image so [music] it's added to the
- 10:01prompt box. Then, paste in the prompt
- 10:03Claude gave you. In the duration
- 10:05section, reduce it to just 5 seconds and
- 10:07make sure the aspect ratio is the exact
- 10:09one you want. Set the resolution to 720p
- 10:13so you don't waste credits. For this,
- 10:15it's charging us just 23 credits, which
- 10:17is fine. So, I'm going to click
- 10:18generate. And while that one's
- 10:20generating, I'm going to upload a couple
- 10:21more just to compare the results with
- 10:23what OmniFlash can do. All right, after
- 10:25a couple of seconds, we've got our
- 10:27animations from both Higgsfield and
- 10:29OmniFlash. [music]
- 10:30So, let's preview what we got. Let me
- 10:32click on the first video to see what the
- 10:33output looks like.
- 10:39All right, this looks pretty decent.
- 10:41Next one.
- 10:45This looks pretty good. Matches the
- 10:47animation style we're going for.
- 10:50Let's preview what Flow was able to
- 10:52create.
- 10:54All right, honestly, the output from
- 10:55OmniFlash looks a bit better to me. I
- 10:58don't know what you think, but I like
- 11:00the OmniFlash output. It looks much more
- 11:02animated. Yeah, I think OmniFlash did a
- 11:04wonderful job compared to the rest of
- 11:05the models because it was even able to
- 11:07create this scene that starts out with a
- 11:09photo.
- 11:10OmniFlash added in that photo of the
- 11:12next soldier who wasn't even in the end
- 11:13frame. It added it in on its own, which
- 11:16is really, really cool. So, if I click
- 11:18on it, everything comes up real quick
- 11:20and looks real good. Yeah, it's really
- 11:22cool. So, all I've got to do now is
- 11:24repeat this process over and over. Drag
- 11:27the next image onto the prompt box,
- 11:29>> [music]
- 11:29>> paste in my prompt, send it, and rinse
- 11:31and repeat for every image until I've
- 11:33turned every single frame into a video.
- 11:36Shouldn't take very long. I think within
- 11:385 minutes you should be done with this.
- 11:40All right, once you're done with the
- 11:41video creation, the next thing you need
- 11:42to do is generate your voice-overs. Now,
- 11:45there are a couple of options. The first
- 11:47is Microsoft Clipchamp. Clipchamp gives
- 11:49you unlimited voice-over generations.
- 11:51It's actually a video editor, but it's
- 11:53got a really good AI voice generator
- 11:55built in. So, if you want unlimited,
- 11:57just install the software and you can
- 11:59generate as many voice-overs [music] as
- 12:00you want. Another good option is CapCut
- 12:02TTS, which gives you access to over 200
- 12:05AI voices inside CapCut. So, if you've
- 12:08got the mobile or desktop version, you
- 12:10can create really high-quality unlimited
- 12:12AI audio. And for a more web-based
- 12:15platform, you could use Noise AI. Noise
- 12:18AI is pretty good. However, the quality
- 12:21isn't the best.
- 12:22>> [music]
- 12:22>> It's not my go-to, but it does have some
- 12:24good voices, and they give you about
- 12:262,000 free credits every single day. So,
- 12:28those are my three free ish voice-over
- 12:30options. Noise AI is limited since it
- 12:33only gives you 2,000 credits, but
- 12:35Clipchamp and CapCut are unlimited, so
- 12:37check them out. But, my go-to, the one I
- 12:40love using the most, is 11 Labs. 11 Labs
- 12:44gives you high-quality AI voices that
- 12:46even let you add emotional tags, which
- 12:49means you can monetize the audio you get
- 12:51from it. YouTube doesn't like glitchy or
- 12:53weird robotic-sounding AI voices. So, if
- 12:56you want to monetize your content, I'd
- 12:57always advise you use 11 [music] Labs
- 13:00because it'll stand the test of time.
- 13:02So, just head over to 11 Labs, sign in
- 13:04with your Google account, and they give
- 13:06you 10,000 free credits a month. Once
- 13:08you're in, click on the voices option to
- 13:09pick whatever voice you want to use.
- 13:11Now, what I like to do is add square
- 13:13brackets and put in an emotional tag.
- 13:15Let's say, whisper. Then, close the
- 13:18bracket. It turns purple, which means
- 13:20whatever I generate in there is going to
- 13:21have that emotional tag influencing how
- 13:23the voice-over sounds. So, I'm going to
- 13:25head back to Claude,
- 13:27scroll all the way up to where it gave
- 13:30me my 1-minute script,
- 13:33and copy just the first chunk. I'll
- 13:35paste it into 11 Labs and generate it to
- 13:37hear how it sounds. If it sounds good,
- 13:40I'll download it, then head back to
- 13:41Claude to copy the next chunk. And I'll
- 13:43just keep copying chunk by chunk until
- 13:45I've generated every single part of my
- 13:47script and turned it into a clean,
- 13:49high-quality sounding voice-over. Once
- 13:51I'm done with that, I need a good
- 13:53platform for music generation. For this
- 13:56tutorial, I'm going to use Suno AI. Just
- 13:58head over to your browser, search for
- 14:00Suno AI, and click the first link you
- 14:02see.
- 14:03That'll take you to their website, which
- 14:05looks something [music] like this. Suno
- 14:07AI is an AI music generation platform.
- 14:09You give it text, and it creates
- 14:11high-quality, monetizable AI tracks that
- 14:14you don't have to worry about copyright
- 14:15issues with.
- 14:16Once you sign up, click on create. Over
- 14:19here, you can create any kind of song. I
- 14:22always advise you toggle on the
- 14:23instrumental button when you're
- 14:24describing a background track. For me,
- 14:26I'm going to type something like silent,
- 14:29growing, intense,
- 14:32driven,
- 14:33blockbuster audio, and then just click
- 14:35generate. [music]
- 14:36Suno takes that simple prompt and
- 14:38creates a specific background track for
- 14:40me. The reason I'm going with this
- 14:42specific track is because I want the
- 14:44audio to be subtle in the background and
- 14:46fit the style of video I'm making.
- 14:48Whatever you're making, you can write a
- 14:49prompt that suits it and generate your
- 14:52track. And of course, you could use
- 14:53Claude or ChatGPT to help you with the
- 14:55prompt. It's all up to you. Once Suno
- 14:57generates your track, all you've got to
- 14:59do is click on the three dots at the
- 15:01bottom left, then click download, and
- 15:04download the MP3 so it saves to your
- 15:06computer. Once you've done that, we now
- 15:08need an editing software to compile
- 15:10everything together into [music] our
- 15:12final published video. For this, we're
- 15:14going to use CapCut. It's very
- 15:16beginner-friendly, so if you've never
- 15:17used it before, don't worry. It's really
- 15:20easy to understand. All you've got to do
- 15:21is create a new project and import the
- 15:23folder where you stored all your assets.
- 15:25For me, I like to keep all my assets in
- 15:27a single folder, my voiceover, my
- 15:29background music, and all my footage. I
- 15:31put it all in one folder, then click
- 15:33import to add it to CapCut. Once the
- 15:36folder's added, I'll first drag my
- 15:37background track onto the timeline, crop
- 15:40it, and reduce its volume so it's not
- 15:42too loud. Then I'll use the magnifying
- 15:43glass to stretch the timeline so I can
- 15:45see clearly.
- 15:46And I'll start dragging my voiceover
- 15:48onto the timeline in the exact order I
- 15:50generated it. Remember, we generated in
- 15:53multiple batches. So, depending on how
- 15:56long your script is, your voiceover
- 15:58might be in a couple batches or more.
- 16:00So, I'll align the voiceovers, then crop
- 16:02the background track to make sure it's
- 16:03lined up and synced to the voiceover.
- 16:06Then I'll select both voiceovers by
- 16:08holding command or control and clicking
- 16:10on both, and I'll bump the volume up to
- 16:13around 6.7 decibels. After that, I'll
- 16:16start dragging in all my clips in the
- 16:17exact order I generated them and just
- 16:19make sure everything's synced up. All
- 16:21you've got to do is drag the clips and
- 16:23line them up properly. For some of them,
- 16:25you might need to crop a bit, but I'm
- 16:26pretty sure the animation software has
- 16:28done [music] most of the work for you.
- 16:29So, just make sure you crop the footage
- 16:31to match the narration being said on
- 16:33screen. Once all your clips are in and
- 16:36synced properly to the voiceover, the
- 16:37next thing to do is start adding
- 16:39transitions between each clip. Now,
- 16:41based on the kind of video we're making,
- 16:43transitions need to be subtle or fit the
- 16:45specific style. So, just go through all
- 16:47the transition styles and pick whichever
- 16:49works best. The camera one might look
- 16:51pretty good. I'd suggest adding
- 16:53different transitions in different
- 16:54places so it doesn't get too repetitive.
- 16:57Just make sure they fit the style or
- 16:58animation you're going with.
- 17:00For me, the ones I used the most were
- 17:02the slide style transitions. [music]
- 17:05So, add all your transitions, and if you
- 17:07want captions, you can add those, too.
- 17:10But I don't really like adding them for
- 17:11videos like this, since the visuals
- 17:13already have a lot of text pointers. So,
- 17:15I'm going to preview it to see how it
- 17:17looks and once it's perfect, I'll click
- 17:18[music] export and export it in 1080p,
- 17:21which is what I exported off flow. But,
- 17:23if you exported your quality in 720p off
- 17:26flow, then export that same 720p here.
- 17:29You could use an upscaler to bump up the
- 17:31quality later or even do it right here
- 17:33in CapCut. But, that's pretty much it.
- 17:35Check out how the video looks.
- 17:37>> May 1925,
- 17:39Paris.
- 17:41A well-dressed man reads a newspaper
- 17:43story about the Eiffel Tower rusting and
- 17:45costing a fortune to maintain. His name
- 17:49is Victor Lustig.
- 17:51He has an idea.
- 17:53Days later, forged government stationery
- 17:55in hand, he invites five scrap metal
- 17:58dealers to a private suite at the Hotel
- 18:00de Crillon.
- 18:02He calls himself a deputy director of
- 18:05the Ministry of Posts. The city, he
- 18:08explains, can no longer afford the
- 18:10tower. It will be sold for scrap,
- 18:12quietly, to avoid public outcry.
- 18:15He picks the most eager dealer, a man
- 18:18named Andre Poisson.
- 18:20He takes the payment and a bribe on top
- 18:22of it.
- 18:23Then he boards a train to Vienna.
- 18:25Poisson, too ashamed to admit he was
- 18:28fooled, never goes to the police. So, 1
- 18:31month later, Lustig returns to Paris.
- 18:35He gathers a new group of dealers. He
- 18:37sells the tower again.
About this transcript
This page contains the full transcript of FREE Claude Prompt That Creates VOX Style Videos AUTOMATICALLY | Here's How by Jacksons AI, generated from the public captions YouTube serves with the video. The transcript has 3,695 words across 551 segments, with the original timestamps preserved so you can click any line to jump to that moment in the embedded player.
What you can do with it
Use the transcript to take notes, quote the speaker, build a study guide, generate a summary with ChatGPT or Claude via the YouTube Summary tool, or export it as a timed subtitle file with YouTube to SRT. You can also re-open it in the transcriber to translate the transcript into 100+ languages.
Free YouTube transcript tool
YouTube2Text is a free YouTube transcript generator — no signup, no daily limit. Paste any YouTube link and get the full transcript instantly, with timestamps, click-to-jump, translation to 100+ languages, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.