I Tested 100+ AI Tools, These 9 Will Blow Up Your Content

0:00Today we're talking about AI tools for video content creation. If you want to learn how to make videos faster, I'm going to break down all the tools you should be using to speed up your content workflow. Now, here's the truth. When it comes to AI tools, there are thousands of them out there, and it can be very overwhelming to know which ones are actually good, when to use them for what scenarios, and most importantly, how to use them, so they actually save you time. So, in this video, I'm going to break down the best ones. These are the nine best AI tools that every content maker should be using and exactly how to

0:30use them. By the way, I'm Callaway. I have a million followers. I've done billions of views. And content is all I do all day long. Okay. AI tool number one for video content creation is ChatGBT. But this isn't ChatGBT in the way you would normally use it. And if you haven't tried what I'm about to show you before, this is going to blow your mind. Now, there are dozens of ways you can try to use chat GBT in the content workflow to speed things up. But specifically, I'm going to walk through two of them using the 03 model. If you haven't tried the 03 model in chat GBT,

0:57this is their latest reasoning model. And it was built using chain of thought, which basically means you give it a task and then it thinks for a while, performs a bunch of micro steps, which it chains together, and then ultimately gives you the answer. It's the closest thing to human level reasoning and thinking that's available on the market. Now, the first way I've been using Chad GBT with 03 is to rapidly find B-roll to give to my editors. So, for example, in this case that I'm showing, I was making a short form video about OpenAI acquiring Johnny Ives design startup. That was a common AI story in the last couple

1:27weeks. And I wanted to give my editor a bunch of links to old Apple videos and iPhone design footage that he could use in the video, but I didn't want to have to go to YouTube and manually find all those links. And I wasn't sure if he was going to be able to find them on his own. And this is a very common problem

1:39that happens in the content workflow. So, I just prompted chat GBT with the 03 model turned on with this prompt. I'm writing a short form video script. I have the script done. I now need to find a bunch of YouTube video source links that I can give to my editor so they can best visualize the lines in the script. And when I say visualize, what I mean is that a viewer needs to hear my words and see the visuals and then have increased comprehension of what I'm saying. So, for example, if I say this guy designed the original iPhone, I would think to have a visual of Johnny IV with Steve

2:07Jobs from 2006 in Apple's headquarters working on it. Can you find me 10 source links of YouTube videos that would have various B-roll clips for us to use for the following script? And then I pasted my script in below. That's all I gave it. I didn't give it anything else. Now, right here, you can see what Chat GBT gave me back. It doesn't have vision capability. So, it's not like Chad GBT can go watch the YouTube videos, identify certain clips based on AI vision, but it is able to source these videos, look at the transcripts, and then it gave me back this insane table with 10 links mapping to every line in

2:40my script, the link, and what potential clip spot in that link I could maybe use. It's pretty insane. And that was one shot first prompt out of Chad GBT3. Now, if you want to take this a step further, and I'm not going to cover this in this video, you could probably build a workflow and either make.com or nadn where you take these structured outputs, download them, clip them up based on the timestamps, and then autosave them into Google Drive folders. And so, you don't

3:05have to go with that advanced level. We're not doing that yet, but just this as a way to speedun getting the B-roll sources for my editor saves him 30 to 60 minutes per video. Now, if you want to take this a step further, which I recommend, you want to create a project in Chat GBT that has these same instructions that will run over and over and over for every video. The difference between a new chat, which is what I showed you, and a new project is that for projects, you can store system instructions that apply like a filter to your prompt every single time you use

3:34it. So, what I would do is take that prompt I just walked through, add that in as system instructions, and then every time I have a new video, I just paste the script in, and it runs over and over and over. You can then hone these instructions over time to improve it or give you a different format. Okay. Now, the second way that I use chatbt that I want to cover here is with the 03 model, but also with deep research turned on. And if you haven't ever used deep research, prepare to have your mind blown again. Deep research is essentially like having a superpowered research analyst at your fingertips

4:01whenever you want. So, while 03 is a thinking model for individual tasks, Deep Research will go out and canvas the entire internet looking for hundreds of different sources on a topic and put together a full research report for you after 5 to 10 minutes of thinking. If you were going to research any topic and really go deep, you would spend hours watching YouTube videos and finding articles, curating all your notes, and then organizing your thoughts. Deep research does this in like 5 to 10 minutes. So when you combine the 03 reasoning model with deep research, you can basically go extremely deep on any

4:33topic in a very short period of time. And this of course is super useful in many ways. For one, you can essentially ingest every article or video on a topic so that you can quickly form an opinion about what type of content you want to make about it. So let's say you wanted to make a YouTube video about psychology tactics in marketing and you had some in the back of your mind, but you kind of wanted to do additional research to see if you're missing anything. You just turn on 03 with deep research, ask it to run that report, which I'll show you in a second, and all of a sudden, you have

5:01the most comprehensive possible report on psychology tactics. Now, in addition to research in the sense I just described, it's also able to do deep analysis on anything you want, which is again super valuable. So, for example, let's say you wanted to find new video ideas that you could make on your channel. Typically, AI would be really bad at this, but with 03 and deep research, it's actually insane how good it could be. So, if I were doing this for my channel about content strategy and social media, this is what I would prompt to find new ideas for YouTube

5:29videos. Let me walk you through this. I'm a creator that makes videos about content strategy, social media growth, and YouTube. Here's my channel, and then I link the channel. I'm looking for a list of new video ideas to cover. Typically, the best ideas are targeting the biggest unmet needs and/or pain points that my target audience has. My target audience is business owners that are trying to learn how to make better video content to grow faster on social media and YouTube to ultimately drive that free organic traffic to their products and services. You can see on my channel which of these videos have outperformed relative to the others. Can

5:58you do a full deep dive on my channels and others in the category and give me 10 new ideas for videos I should make? For each one, can you break down the idea, the title, and a description for why you think this idea would do well for me? Feel free to ask me any follow-up questions before you run the full process. So, here again, I'm using deep research to help me come up with new ideas for my channel. Now, I'm not going to go super deep into the exact list it gave. I'll scroll through it quickly so you can see, but I recommend taking that exact prompt and trying to

6:24use it for yourself. The one thing I did want to call out, and this is the pro tip for AI, if you take nothing away from this video besides this, you will be in great shape. When you prompt deep research or any AI tool, feel free to ask me any follow-up questions before you run the process. So, for example, in this case, I asked that and one of the questions it asked me were, can you link competitor channels or other channels in the niche for me to compare to and so I went out and I added four or five other channels and then the results are so

6:50much better. Now, you can access this level of 03 and deep research in the $20 a month chatbt pro plan. This combo 03 and deep research is my favorite AI tool to use. All right, AI tool number two for video content creation is sandcastles.ai AI and this is my tool for short form video script writing. If you make videos on Instagram reels, Tik Tok, LinkedIn shorts or YouTube shorts, Sandcastles is the gold standard in script writing. And on Sandcastles, you can either paste in a draft of a script and we'll punch it up. You could paste in an outline, so notes, and we'll only use those notes to write the script, or

7:24you could paste in literally a single line, and we'll go do the research and then develop the script for you. You have those three options. The Sand Castle scripts are way better out of the box than something like Chat GPT or Claude. And that's because we've spent a ton of time building in advanced logic and prompting based on everything I know

7:41about storytelling and script writing. There are not a ton of video content creation specific AI tools on the market, and that's because most tools raise a ton of VC money and have to broaden their market that they're going after. So, they become more generalized, add a bunch more features you don't need, and become bloated. But Sand Castles was built specifically for video content creation, in this case, script writing. Another thing that makes Sand Castles really powerful is that we built a dedicated hooks engine. So when you're writing your script, there's a module where you can swap in one of four different hooks that have all been

8:12validated based on story structures that we know work for short form. And then once you generate the first version of your script, you can iterate it over and over and over with a bunch of different tools. And we show you this AB visualization that maps to what's changing from the original version. So again, if you make short form videos and you struggle with script writing, you want better hooks, you just want more compelling stories, Sand Castles is the gold standard and will be the easiest for you to use because we've designed it for this use case specifically. And if you're watching this video, you can try

8:38it free at sandcastles.ai. All right, tool number three is called Sort Feed. And this thing is super simple, but actually really effective for social media content ideas. The most effective way to find viral content ideas is to look at other creators in your space and try to find commonalities across their hooks, their visual structures, their story structures or their ideas. So essentially find what worked and then remix it for you. That's the strategy on YouTube, social media. That's what works everywhere. Now the manual way to do this is by having a list of creators in your niche. Every day you open up, you click the link, you manually scroll

9:10through, you look at the metrics by hand, you try to compare to what a baseline is, you find the outliers. It's a very manual and timeconsuming process that people have to do every couple days. But Swordfeed is a tool that makes this much much easier. Swordfeed allows you to go to any Instagram or Tik Tok account. You pick the Chrome extension, you activate it, and then it allows you to resort the profile of videos by views, likes, comments, and you can do it across a date range or a specific number of videos historically. It's kind of like this magic layer that sits on top of what's actually there on the

9:40social platform. And this makes it super easy to analyze outlier performers, the best performing videos from your competitors without having to guess which ones are better. So you can see I just ran it on my own account and I selected I want to analyze the last 100 videos and I want to sort by top views. I just run it and after a couple seconds immediately it's resorted my videos in that grid in order of top views. So now I can open every video in a separate tab and I can quickly analyze the idea. Is there a commonality in the hook? Is there a commonality in the visual

10:09structure? I can really start to analyze patterns much easier because I have this list. If you want to try Swordfeed, it's swordfeed.com. I have no affiliation. The free plan allows you to try it. I think there's a $40 onetime license that allows you to get the full access. Now, I will let you in on a little secret. We are dropping something soon inside of Sand Castles that will increase what

10:27you're able to do on Swordfeed by 10x. It will make finding and remixing content ideas way easier. And if you want to be on the beta weight list, you can get first access to that feature. I've got a link below that you can sign up for. It's going to blow people's minds. All right, the fourth AI tool is Craya.ai. And this thing is extremely slept on. I don't hear anybody talk about this. One of the most helpful AI use cases in content creation is using AI to generate images and videos that you can use as B-roll. So, this of course means describing a visual that you want to bring to life with text or

10:58taking some visual and then animating it with motion to make it a video. But here's the common problem when it comes to AI images and video tools. There are dozens to choose from and it feels like new models are coming out weekly. Some of the models are international, some are US only. Some of the models can be run with a clean front end. Some of them you have to run some local instance server on the back end to be able to access. It's all over the place and way too much to keep track of if you're not actually in the weeds on AI. And this is

11:22why I like Craya. Craya is a clean front-end experience that pipes almost all of the different AI image and video models into one place. So, you can basically pick from a drop down and then access whichever model you want. So, for example, let's say I was making a video and I wanted a clip of a Red Bull can in the sky in the clouds kind of like floating down and I was envisioning that

11:41clip. So, I wanted to make it with AI. So, first I can pick whatever the image preset that I want and then I can get my prompt kind of dialed in. But the cool thing is in the bottom left I can then open that menu and I can click from any of the different AI image models to try. And in that list, they tell you which ones are fastest, which ones are cheapest. You can kind of pick a couple different ones to try or just go for your favorite. And what's cool is this allows me to rapidly generate a bunch of different image options from different

12:06models because they all differ here or there. They're all better or worse at certain things. Now, being able to hot swap between the models is an amazing UX perk. But that's only half the value of Craya because then once you have the AI generated image that you want, you can click that little video icon, tab over to the video feed, and run the same process to bring that image to life. And again, you've got a list in the bottom left of all the latest video models you can choose from as well. So, just like that, in a matter of 30 seconds, I've generated a bunch of different image

12:32choices and then immediately prompted to the video models that I want to animate with. Right now, Cling 2.0 is my absolute favorite video model, but it's super expensive and super slow. So, I'll often use WAN 2.1 or Runway. This endto-end flow for creating AI assets, text to image to video, all in the same place, is a super powerful workflow and can really replace almost all of the stock footage and stock imagery that you're using. Now, if you're serious about content, let me give you a bonus tip on this as well. The bonus tip is that if you have an editing team, you can actually create a collaborative

13:04workspace on Craya. Add all of them to it and then you can see collaboratively who uses which prompt to make the best image, to make the best video, and also buy compute tokens that they all can use independently. So, for me, the future workflow that I want is having all of my editors have access to the same workspace. Everybody can experiment and then when we find something that works, we can all take that prompt and reverse engineer it. That's that iterative AI process that you want to build. So you can play with everything I talked about in Craya at Craya. All right. Now, before I keep going, I just want to

13:34mention if you're a business owner and you're serious about leveling up your content, short form, long form, on YouTube, whatever the video content is, you want to get better at video, I built a free community for business owners called Wavy World. We have over 21,000 other creators, entrepreneurs, and business owners in there that are all working together collaboratively to improve their content. And I put tons of free trainings and tutorials in there for anybody to access. So, if you want an invite link, it's free. You can check the description. All right, let's keep going. Now, there are two other AI image/v video tools that are not yet

14:06available on Craya. So, while Craya can serve almost all of your creative needs, I recommend gut-checking Craya against these other two tools as well. You likely don't need these, but these are kind of my favorite and the best in class each of them. So, I recommend kind of gut-eing between them. All right. Now, the first one, AI tool number five, is brand new, and this is Google Veo 3 with Flow and Flow TV. Let me break down what all those words mean. Veo 3 is

14:31Google's third generation video model. So, you put text in and you get video out. And the reason this is worth playing with is because the outputs are actually insane. The quality is so good. It's indistinguishable from real footage. And the videos come out of the box generated with audio. So character dialogue and sound effects, background noises, all baked into the video. It's crazy. Now, you have two ways to use Veo 3. You can either use a lightweight preview version that only allows you to make like 8second clips via Gemini, and I'll link that below. So, it's like Google's chatbt competitor. You can kind of just generate videos directly in

15:04there, or you can use Veo3 in Flow, which is how they want you to use it. Now, Flow is marketed as an AI film making tool from Google. So, it combines Veo 3 plus Google's latest image model all together and creates this really interesting workflow for building short films and AI type videos. And Google actually worked with a bunch of super

15:25serious filmmakers to design this tool. So, it has a bunch of unique features that you don't find anywhere else. The main one, which is really cool, is called ingredient mode. So essentially, let's say you generated an AI video of a woman in a red dress getting out of a yellow cab in a nighttime background in New York City. What you can do with ingredient mode is actually extract each of those separate ingredients. The woman, the red dress, the yellow cab, New York City, and then you can hold those pieces constant through other clips. This is how you create scene continuity and consistency. So if you want that same yellow cab with that same

15:57color of yellow showing up here, here, here, here in all these scenes, you can do that with ingredient mode. Now, the last piece of what Google released, which I think is super underrated, nobody's talking about, is Flow TV. So, like I said before, the limiting factor of these tools is not the model power, it's people knowing how to prompt to get what you want out of it. So, Flow TV is this free resource which is basically a pseudo streaming service except all of the channels are AI video generated via Flow and you can flick through the channels and also different videos within those channels and then toggle a

16:29switch to reveal the prompt that was used to make them. So, this is kind of like a Pinterest forever scroll but for AI video and it can be really helpful to get inspiration for how to prompt. Now, I believe Craya is going to add Veo3 via API into it soon because it already has Veo 2, but they're probably not going to be able to add Flow into Craya. So, if you like the UX of Flow, you should use Google Veo3 in Flow. If you like the UX of Craya, you should use Craya and just assume Veo3 will actually come. This is what I mean by the UX preferences. You

17:00kind of have to pick the workflow that you like the best. Now, you can play with Veo3 in Flow on a limited basis on Google's $20 a month plan. Again, they released AI Ultra, which costs 250 bucks a month, which is insane, but that's how you get ingredient mode and unlimited features. It's kind of annoying these AI companies are coming out with these massively expensive models, but that's what it is. All right, AI tool number six, and my personal favorite for generating AI images, is MidJourney. So, for MidJourney, you put text in and you get highquality images out. Now, if MidJourney was inside of Craya, then

17:29Craya would be the only tool you need, but sadly, it isn't. So, this is what I was saying. Craya solves most use cases, but Veo 3 with Flow and Midjourney are the two that aren't in Craya, which is a bit annoying. But Midjourney is by far the gold standard for AI image creation. And so this is a potential workflow for you as a content maker. You go to Midjourney, you prompt with text to get the image out, you take that image into Craya or Veo3, and then you bring that to life. So midjourney to Craya, midjourney to Veo 3. Those would be the two workflows that I'd recommend for

17:59going from text to image to video. if you feel like the midjourney images are way better than what you're getting from the Craya models. Now, the reason why I like MidJourney so much, as soon as you use it, you'll realize there's kind of a different feel to it. And this is because it feels like MidJourney prioritizes aesthetics over everything else. Essentially, it was built to generate beautiful imagery and then all the rest of the bells and whistles and speed and workflow came on top. That's the reverse of how a lot of these other models were built. Most models were built for speed and cost and then the

18:26imagery was third. Midjourney also has a really cool style ingestion process. So, when you want to use their latest model, V7, it'll show you 200 different AB combination of images that you go through and rate. What this is doing is honing your unique style preferences for the types of imagery across many categories that you like. And so, what that does for you is when MidJourney generates an image for you, it's almost like it's a third draft instead of a first draft because they already have

18:51all of your style preferences built in. Now I will say midjourney's one issue is that it's not as good at rendering text on the images. So if you want to build something like an infographic or a poster and you want that text to be pristine, you want to use chat GBT 40 image which is either in Craya or you

19:07can use it directly inside of ChatGBT. But for visual aesthetics and cinematics, Midjourney is the absolute best. Again, to get good at midjourney, it's an art and it's honestly probably one of the most valuable skills you could invest in over the next 3 to 5 years. And there are a few free trainings specific to midjourney that I've linked in that doc which is all below. So to summarize the workflow when it comes to text to image to video that I would recommend. Either you do everything in Craya to simplify or you start with midjourney or chatbt 40 image if you need text. Take that image out

19:37and then go into Craya or go into Veo3. Google Veo3 to animate. All right. Tool number seven that I really like using is Claude, but specifically Claude projects. Claude projects work similar to chatgbt projects in the sense that you can store system instructions that apply like a filter to anything you prompt it. But the reason I like claude instead of chatgbt for a specific use case is that claude 3.7 sonnet and now for sonnet are just way better at creative writing than chat gbt. GBT is better at research. Claude is better at writing. Now if you're absolutely cashstrapped, you can probably get away with just paying for either Chad GBT or

20:11Claude and kind of forcing the research and writing in one or the other. But if you have the money, I recommend having both because they're kind of better at different things. So, what exactly do I use Claude projects for? When I'm writing scripts for videos, I go back and forth between Sand Castles and Claude. Sand Castle is actually built on top of Claude. And like I said, it's very fine-tuned to work extremely well out of the box, but there are a couple features we don't yet have in Sand Castles that are coming soon that Claude has that I really like using. And so, I kind of go back and forth between them

20:42right now. One of those features in Claude that I love is rapidly updating system instructions. So in a Claude project, I could paste in a script or I could generate a script. I could re-edit it by hand. Then I could ask Claude to analyze what are the differences between my new script and the first version. It will break out the differences. And then I could say, "Okay, update your system instructions to include those differences moving forward." And so what I'm doing is recursively improving those instructions over time. The more manual

21:10training I give it, the better it gets. And this means you can hone it with much more personalization versus what you can do in sand castles today. Now, we're building that into sand castles, which will make it all great in one place. But until we do, I kind of like to go back and forth because here's the issue. If you use Claude without sand castles and you're trying to write scripts, what that means that you have to know what system instructions to give it to get it to generate a good script. And then you would have to know how to manually tweak that script so that you can give it

21:36something good to recursively update. If you don't know how to write scripts, well, you're not going to be able to do either of those things. The beauty of Sand Castles is we have built that into the infrastructure. So, you don't have to know what good is. We know what good is. That's why using Sandcastles for

21:50script writing is better than Claude. So, right now, my workflow to make a video idea looks something like this. I'll have the idea. I'll go to Chad GBT deep research and I'll get all of the facts on that idea that I can't glean myself. That speedruns the research. Then, I'll go into Sand Castles and I'll write the script. I'll take that written script and either I'll manually edit it or I'll go to the Claude project that I've built, put that in, iterate the script, do another manual tweak, and then update the system instructions in Claude. So, I'm stitching these things together because I'm trying to build the

22:21best workflow that I can then share back with you guys. If you're a beginner, I recommend just using Sandcastles out of the box. You don't have to worry about system instructions. You don't have to worry about how to prompt Claude. Everything is built already for you. But if you're more advanced, that's kind of the workflow that I'm doing right now until we have all those features built into Sand Castles. Okay, the last two tools go hand in hand together. This is number eight and nine. So, I'm going to talk about them together. And I'm not personally using these two tools in the capacity I'm going to talk about, but I

22:46know a few creators that are, and it's really crushing. Tools number eight and nine are Hey Genen for AI avatars and 11 Labs for professional voice cloning. Now, there are three channels that I want you to look at that are using Hey Gen AI avatars. Rowan Verun and Nick Serev. So in 11 Labs, you can create this professional AI voice clone where you upload 30 minutes to three hours of audio, clean audio, and it will create a voice that sounds identical to you. You can hear this is the one of me that we've built that sounds really expressive. If you know my short form videos, it sounds just like me. Then you

23:20go to hen and you build an AI avatar, which is the video. Hen has audio also included, but you can just API the 11 Labs in because it's way better. So, you use Hey Gen for the video, you connect 11 Labs inside of Hey Genen with your professional voice for the audio and all of a sudden you feed it a script and it sounds and looks pretty compelling. Now, it doesn't look perfect. If you just stare at the avatar, the movements are a little robotic, the mouth is a little off, but if you cut around it like these guys do, you can't really tell the difference. This workflow would not work

23:48for long form YouTube if you wanted to be a face creator because you'd just be able to tell, but it's very passible for short form video on social media. Now, the reason this is compelling is that it's already working for these guys and they're producing a ton of volume using this approach. They don't actually have to sit down to make the video. If they have somebody writing the scripts or sand castles, they can feed it directly into their AI avatar, which prints the video. And all they need is an editor to

24:12put it together and then post for them. They don't actually have to touch Instagram or Tik Tok. It all happens automatically. Now, the reason I don't use it is because I haven't been able to get my avatar to look convincing enough where I feel like I wouldn't lose trust by using it, but I'm working on it and I will share the full playbook when I get it working. All right. Now, before we go, one last bonus tip on 11 Labs, cuz 11 Labs is kind of like the AI audio king. 11 Labs has a really cool feature called sound effects. You can click sound effects, describe any sound effect

24:39you want, and then 11 Labs will print you four versions of that sound effect from scratch. This feature will be everywhere, but I really like the UX that 11 Labs has. So, if you're trying to up your audio, that's something that I use 11 Labs for. All right, and there you have it. Those are the nine AI tools that I'm using and that all content makers should be using to speed up their video content workflow. And all of the links to all the tools I referenced, all the training docs, everything is going

25:04to be linked below, organized for you. If I have a discount code, I'll put it down there for you so you can try these tools for free or as cheap as possible. And here's one final summary breakdown of everything I mentioned, but layered into the workflow for content creation, right? All right. So for content, you've got ideas, research, script writing, recording, and editing. Those are kind of the buckets. So for ideas, I'm using sort feed to do some of that work to kind of organize the reels and find patterns. And then I'm using ChadBt 03

25:29to actually come up with fresh ideas. I'm combining those things. For research, I'm using that combo of Chadbt03 plus deep research, which is the king combo. For script writing, I'm using sand castles and clawed projects back and forth. If you don't know what you're doing, just use sand castles. If you do know what you're doing, I recommend doing sand castles and then having a clawed project that you train and using both of them. For recording, I'm not using, but people are using Hey Gen plus 11 Labs, what I just went through. And then for editing, you've got Craya, which does image and video, Google Veo 3 and Flow, which is video,

25:59MidJourney, which is image, and 11 Labs, which is sound effects. And I went through a couple different combinations of workflows that you could try with those. Again, right now, my life is firmly planned at this intersection, AI tools and content strategy. My only goal is to build the best tools, the best systems, the best information to help you figure out how to use content to grow your business and make AI workflows that allow you to make content easier, faster, better. So, if you want me to make more videos like this that are a little bit more focused on AI instead of the theoretical content stuff, please

26:29leave a comment and let me know and I can start doing that. And remember, ton of free stuff in the description. Just check it out. There's so many links. I'm just giving away the answers. If you have this problem, check it out in the description. All right, we'll see you guys on the next one. Peace.