NEW AI Video Tools in 2025 (Must Know!)
I've dived through hundreds of new AI video tools to tell you which ones you shouldn't Miss there are a whole host of new and exciting tools that we get to get our hands on everything from free AI video generators to some incredible AI video effects that we've never seen before so without further Ado let's jump in first up we have an incredible free open-source AI video model from Alibaba called wanks now it's an unusual name now in popular benchmarks wanks is out on top let's take a look at some of the examples coming out of this new AI video model so here we can see an incredibly complicated and difficult situation to render with AI video and wanks is performing absolutely incredibly you can see how consistently this ice skater moves in a way that makes it believable now this is an incredibly challenging situation to render effectively the only thing you would say is odd is at one point the legs seem to switch if you watch very closely at this point now there is a morphing of the legs now we have another demonstration video for us to take a look at this is the one 2.1 model and it comes in four different sizes but the impressive thing about this is its ability to handle complex human movements often elements like dancing or fighting are some of the most challenging things for AI video models to render out effectively and you can see here that one EGS is creating some of the most liike interactions that we've seen from an AI video model there is a whole host of impressive examples and a real degree of cinematic quality coming out let's just pause the video here and dissect these with a little bit more intent so what you can see is this Catman this anthropomorphic cat in a very stylish Urban attire is looking like he's coming home after a long day's work in the office trying to get home before his child's bedtime his little kitten is wrapped up in bed and here we can see the wonderful shades on this saxophone Edition really using a monochromatic approach that gives it a lovely stylistic intent and just see how beautiful and anatomically correct this woman is it's a real step forward in the quality and Fidelity and realism of AI video now it's able to create videos in different styles as well not just in photo realism we can see a number of different animated Styles here there is a little fox looking rather mischievous and also this 3D animation is beautiful now it's also able to work with text which opens up a whole world of possibilities for creating Motion Graphics now the great thing about this model is that it's open source which means you can download it install it and run it on your machine however it is fairly slow to run if you use it locally on your own device but the good thing is is that we can run it in the cloud ourselves so one of the easiest ways to use this model straight away is to go to this replicate link I'll leave a link to that in the description below and you can handle it yourself so here I have uploaded an image myself and this is the video that I got out and you can see that she moves in a very smooth and gentle way now this model is extremely easy to use if you're using it in replicate you can start off by uploading an image and adding in a prompt you then have the option to change the orientation of the video and you can make it either landscape or vertical you can also select the frames per second from 5 to 24 and you have the option to choose a balanced or fast mode balanced slows it down a little bit but gets you a better result now the pricing for using this model on replicate is pretty expensive it's actually 50 cents per second video now you are able to run this locally but it can take up to 4 minutes to get a generation and that is just for a 5-second 480p video so this is a huge step forward in open source AI video models however it's still rather slow and expensive to run in the cloud but an incredibly impressive model and what we want to note most of all is the ability for this model to work with complex movements here's another example example of a beautiful Dragon here we have a woman who is drinking a cup of coffee and the way that the behavior is animated is incredibly beautiful the only thing I would comment on on this one is just watching this arm there is a bit of morphing as it goes behind the table it suddenly shrinks down to a little claw and then returns to a more normal size now if we look here in this example how well the fire is rendered is incredibly lifelike in and this is a very moving and quite disturbing scene even and here we can see a little bit of a inclination towards how the video model can handle action and you see that she is able to wield her swords in a way that does not Bell reality and I particularly admire how the light refracts on her cheek here as she brings up her glowing lightsaber now this model is made by Alibaba it's free it's open source and there are four different models available each one is of a different size and speed offering you different resolution and different quality obviously the smaller models are faster to run and the larger models take a lot more compute now we are in a real period of luxurious opportunity with the AI video models that are available because Google has recently launched its V2 model and this is another fantastic AI video model to keep your eyes on let's take a look at a few of their examples here you can see some stunning examples this model creates very realistic videos however there is a not a high degree of dynamic to their motion you can see all of the examples that they've cherry-picked here have very limited amounts of action or movement or camera movement now you can create videos up to 4K in resolution with vo which is extremely impressive now it appears to maintain great coherence in these models you can see here that this woman continues to look like the same woman as we progress through the clip and I would have to criticize her dancing a little bit it's a little bit odd but apart from that it is beautiful and quite an aesthetic shot as well certainly with this out of focus bands in the foreground uh there's certainly a creative intention here now let's take a closer look at this example from VO2 and this is extremely important to observe because what we're looking at here is that when this honey is drizzled on top of these pancakes what we want to see is an appropriate amount of Honey being left on the pancakes and this shows that the model understands how much liquid is actually existing in the shot and you can see that this feels pretty good however there's a little bit of morphing going on as the honey hits the pancake now we can look at a coffee being poured here and this is exquisite you because you would have to say that the coffee is being filled up at exactly the rate that you would expect it to as the liquid is being poured into the cup perhaps I would say it's feeling a little bit fast for the amount of flow that we see but the way that the liquid is undulating across the surface is absolutely stunning now this is where I think the model falls down a little bit compared to other models that I have looked at you can see that this High paste action shot has a sense that it is not real that the car is moving unnaturally it is doesn't even feel like it is connected to the TAC you can see that it's almost floating over the space and the way that it turns around this corner you can see there is a delay of how the car responds after the wheels have been turned so you can see here that the wheels are turning to the left and there's not much of a reaction from the car however for True realism this model does exceptionally well I think this is a shot that really shows us the strengths of this model it gives us an instance where there is very little that we can pick out that is an obvious artifact of AI video and and the material rendering here is absolutely Exquisite I am blown away by just how natural this synthetic lab coat looks and it catches the light in beautifully refined and detailed way even this this slight texturization on the jacket is a wonderful attention to detail and the fact there are some beautiful minor imperfections that are kept in this shot is another homage to the quality of VO if you just trace this mole above her cheekbone here and that remains consistent throughout the entire shot and that is another thing that can be very very hard for AI video to maintain now it always amuses me that they put these benchmarks on at all of the release websites for any video mod or any AI project and it's incredibly interesting to see that somehow always there is a preference for the model that is is being released I always love the saying that the most important factor in any study being created is who is funding the study so I always Reserve judgment for any type of Benchmark that is not an independent now they do mention that maintaining complete consistency throughout complex scenes or those with complex tion remains a challenge and that certainly matches up with how I interpret this model that the strength of this model is more to do with its degree of realism in slow and graceful shots rather than anything that is too complicated or dynamic here you can see a lot of morphing going on with the characters in the background and whenever you see an object pass behind someone you usually see a slightly new version of the object appear on the other side so you can see with this version that as we come around to the other side there are suddenly more candles and again this is an example of an ice Gator and you see comparing this ice Gator to the one that we looked at just a little bit earlier from One X so if we compare these two side by side it is obvious that 1X is creating a much better rendering of this Dynamic movement however I would say that the general aesthetic is more appealing in the vo version and this is why it's important to learn the capabilities and promise of all AI video models because there will always be situations where one model works better for the thing that you are looking to generate now the pricing of V2 actually Compares quite similarly to the pricing of 1x through using it on their website it costs 50 cents per second a video which is roughly the same as using 1X on replicate however of course if you have a good machine with great vram you can run it locally now the advantage of VO though is that you do get a much better user interface to work with next up is is today's sponsor which is vid AI now vid aai is an AI tool that allows you to create entire pieces of short form content with just text prompts it can create the scripts the music the images and put all together in a seamless process let's take a look at how it works if you come to vid doai you can go to new project and from here you get the option to choose the type of content you want to make we're going to be making a faceless YouTube short so if you want to make a YouTube channel but you don't want to show your face this gives you the tools to do just that next up you can give it a title so we're going to make this project about facts about and Artica next up you can put in a prompt and this will be used to generate the script for your short after this we have the choice of selecting your length I'll go with 30 to 60 seconds and now we get to choose our narrator there are a whole host of realistic voices to choose from but I particularly like Adam he has a well-rounded deep yet authoritative voice now the next part is to choose the type of AI generated images that will accompany your video now you have a number of different styles available here there is this cinematic feel which has almost a ghostly effect and there is also the opportunity to have a more Illustrated comic effect you can also choose from graffiti isometric but my favorite is natural and that's it now you can go ahead and generate your content now this takes less than a minute to come up with a first draft and from here what I love about this tool is that you can then edit different parts of your piece and choose what you would like to change or improve so first of all you get the script and you get to see what images are generated at the same time as this is being spoken you now have the option to regenerate an image if you would like something a little bit different yeah I I like this one a lot more there's a lot more depth going on here now you can also update the text if you feel like something different but I think it's working very well together you also get an expected duration and you can go ahead and click on next as soon as you're happy and then what we get is our beautiful short now we can choose the type of music that we would like to add to our video after that you have the option of adding captions all automatically generated with AI you can select either multiword one word and then the style if you want the Beast style or something a little bit more professional mature and reserved for a fine nature document like this you can also adjust the font size I think I have it a little bit bigger and your own color palette there are also some subtle animations that you can add to the images which include a shake effect and also having a zoom in effect now once you've made all of your decisions you simply go ahead and generate the final version and this renders out your entire short for you and here is how mine came out did you know Antarctica is the coldest place on Earth it holds about 60% of the world's fresh water surprisingly it's also a desert the continent is home to Unique Wildlife there are are no native land mammals here now this thought is great if you want to experiment creating faceless YouTube videos now what is particularly impressive about this is you can quickly automate and scale the process to generate hundreds of shorts in a very short space of time now you're not able to just automate and scale your content production but it gives you the creative freedom to dive into every single detail and perfect it to your own tastes automate your own short creation and try vidai out with a link in the description below and a big thanks to vidai for sponsoring this segment of the video now I also want to reveal to you a new update from Pika because they've just released their 2.2 model and it is better than ever and Pika has really come a long way since it first came onto the scene because it's adding in some incredible effects that differentiate it from other AI video models now their video 2.2 model is an improvement an incremental Improvement on the quality of video coming out inside of their platform but what I want to tell you about is the other features that they have recently introduced because they actually create something that gives us a lot more possibility than other AI video models and these include pea frames which is an incredibly original way that allows you to merge between different shots let's take a look at a creative example from Min Choy on X here here you can see this beautiful fashion edit and what's great about this is it morphs between different key frames in a wonderful artistic and seamless way it does not feel as if it is jarring or being forced it creates this lovely rotating transformation of the model as she switches between different outfits and this is the creative potential of pea frames it allows us to create unbelievable transitions that were never before possible so here is another example and what happens is you put the first frame in and then you put the end frame in now you get to add in a prompt which defines the transition so here we have a smooth transition of a bullet firing from a gun barrel hyper realistic and cinematic and you can see here just how effectively that this transition blends from the start to the end of the shot creating a beautifully mesmerizing incident let's have a good at this ourselves why not so to use peer frames all you have to do is Select pea frames in the tabs at the bottom and then you can choose your first frame so I've been working with some abstract chromatic art styles recently and I think I would like to try some of these out as my frames so what I'll do is I'll go ahead and upload my first frame by selecting the first first frame and choosing I think this Blobby chromatic liquid and then I will add my last frame so I think I would actually like it to morph into this woman or maybe a more lifelike woman I think I'll go with this one let's have blob smoothly trans so I've written in a blob smoothly transitions into face and then I can go ahead and submit that and this is how my video came out and I think it's quite incredible so paa has 2.2 it also has paa frames which allow
us to create very original transitions but it also has this wonderful feature called Pika swaps and this allows us to be incredibly creative with using real life footage and changing elements inside for example I can change myself in this very video into Buxom Boda lady hello and from now on we'll be completing this entire video with me Samantha but you can also turn yourself into a leopard man and that's not all you can actually also use inanimate objects and change those as well so I might change this flask into a torch a flaming torch and I could say it's time for us to explore the caves of Mario now that is Pika swaps where you can swap elements inside of your video but you also have the opportunity to use P additions which is where we can add new things into our video so I might have a nice friend join me for now in this shot and this little guy is entirely created with AI now to use both of these they follow a fairly similar process you can select either peer editions or peer swaps at the bottom of the screen and let's go with peer editions now you upload your video and then you can upload an image and this image will be the reference for adding in an element or swapping out an element so you can create an image in any of your desired image generators I might have this pair upload the image and then you go ahead and write a prompt about how you want to add the element or replace the element in your video finally you can go ahead and submit now one thing I particularly admire about p is the way that it's managed to maintain a a simple intuitive and refined user interface some AI video models have gotten to the point where they're adding so many additional features that they're becoming over complicated and difficult to use whereas paa you can simply jump in and get going whenever you like now the best thing about Pika even better than that is that you can try it out for free you get 80 monthly credits and currently get access to P 1.5 P editions and P swaps but if you want to use the latest features EG the 2.2 model then you're going to upgrade to the $8 per month billing however a big caveat on this which is something I don't like about p is that this subscription still means you cannot download your videos without a watermark and they have quite an infuriating Watermark that doesn't just remain in one place it hops all around and I understand they do this so that they can keep a low tier offering available for everyone but it is a little bit frustrating and you will need to upgrade to the $28 per month plan to get rid of the watermark but I'm pretty impressed with the elements that have been created inside of Pica now here is my addition this is great the the pair is being consumed by the Chrome liquid very enjoyable next up I'd like to show you one of the most advanced AI voice models that I've seen and this is called Hume and the exciting thing about this is that these AI voices understand emotional expression you get to prompt the voice type that you would like so you can really Define exactly the characterist istics of your voice and then offer it a script now let's take a little look at their demo here ye here we can see they've adding the voice prompt medieval town crier projecting his deep booming voice speaking in a Shakespearean British I let it be known throughout the land our model octave doth bring forth speech from text right then a glorious day for a simple folk I suppose and here you can see a medieval peasant with a cocking accent or raspy voice dripping with sarcasm we don't have enough of them Tex speech models already in it octave be not some common text to speech Contraption N is a mighty large language model graced with the gift of voice and what good that so here their demonstration video is also explaining how this differentiates from other text to voice models in that it has the intelligence of an llm in ing the emotional tone and Cadence of the speech which informs how the voices are generated this is because it's a text of speech system that understands what it's saying it's a voice-based llm that means it understands what words mean in Context so it can predict emotions Cadence and more here is another example of this model already be it's built on a great Mighty language model that's that means it knows what meaner in right place can Su outs and now this is certainly what I think to be the most capable and inspiring AI voice generator and is certainly one to check out next up is an interesting way to use your AI tools and this app is called flora and what it is is an intelligent canvas that allows you to combine and use different AI Tools in one place and it's certainly bridging one of the challenges that we have with AI right now that is every day there is a new AI tool to use and to really get the best AI tool for each part of your creative process you'll be using many many different tools and that means you're jumping between different systems you're managing different subscriptions and you're constantly trying to get different apps to work well with each other now that is what Flora is trying to overcome and I think it's doing a very creative and original job at this so how it works is you have this huge infinite art board very similar to a figma board and then what you can do is you can add in different elements then you can link those to other elements for example here you can take three images and put them into a video it will then generate a video based on the three images here in the demonstration you can see the ability to also work in multiplayer with other users so everyone can access the same board and what we can do is we can start talking to the images that we put in so you can ask questions to an llm about the assets that you add into the model and then you can take images you can ask for things like what are the colors in this then you can take those colors and create more videos with other tools that keep a consistent color palette now you also have the opportunity to select what video model to generate your videos with now the ability to generate text image and video all in one artboard uh brings up a much more fluid and open process for creating our work and it's something that certainly appeals to me because you're able to rapidly generate and expand on ideas without having to mentally switch between different platforms now the task switching between different interfaces and different types of media can be a great blockade in truly expressing yourself with AI now currently they have a number of different tools that you can bring in including Halo which is also known as Minimax dream machine which is also known as Luma you can bring in chat gbt and the open AI models you can use cing paa Runway stability AI topaz and recraft and this gives us a lot of opportunities now you can also try this out for free and experiment with making your own layouts now to use it yourself all you have to do is go to new project from here you select what you do in life let's say I'm a creative serviceman and I have a studio and then it gives you a number of different workflows that you can look at you can choose which kind of flow you wouldd like to use and then you have the option to navigate around and add in text image and video as you please you're able to easily connect your blocks and also share this for live editing with your collaborators so here you can see a complex board that has a real visual Network layout so you can see how many different iterations can be chained together to explore a different areas or branches of thought now this is great if if you're really trying to pin down a specific workflow and get everything working in unison in one cohesive way and it's a very original way to approach generating media with AI now speaking of how we can leverage other tools to improve our workflow I have to tell you about Notions new AI feature which allows you to upload images and extract text so this means that you can go ahead head and work naturally on paper mapping out your thoughts and then getting notion to immediately document this in a place for you which is a great way to still maintain the beauty of working offline which I think is also one of the areas that's going to expand and expand in future now particularly one of the features that I think is fantastic about notion is it's AI that formats documents for you now this is something that other AIS really struggle to do well and I use this every single day in notion because what you can do is you can take just the list here and then you can say exactly how you would like it formatted so then you can see it takes it from um standard unstructured list and then what it does is it adds in headers with different font sizes it adds in special types of content for example for here it's turned into a to-do list so that you can tick off these ingredients one by one it's even added in textiles so here the instruction serves 8 to 10 is in italics and it's put it all in a box with these lovely blocked out headers with Emojis now this is great because we can use this process in any type of creative flow that we want and I suddenly see us expanding more into leveraging real world creation types with AI and there is another project that I think is beautifully exciting in this domain so let me show you how this works it is a standard pen that looks unremarkable from the outside it has a little bulbous end which maybe draws your attention but the incredibly beautiful thing about this pen is it contains cameras inside of the end of the pen that tracks everything that you draw so you can see here it has these cameras built in and the reason this is so exciting is that it immediately digitizes everything that that you write down so you not only have a physical copy you also immediately get a digital copy that appears on your phone or on your computer now not only does it keep this as an image it also immediately transcribes it into text and you're able to search through this and record every scribble that you make and I love this approach to technology where we're really integrating the natural feeling of working with pen and paper and bringing it into a digital situation now I think this is a huge trend for AI and what we're going to start to see is this synthesizing of AI content and real content of the real world and the digital world where we're starting to blur the line of where it ends and where it begins about how we can take real life footage content and experiences and also extend them expand them and integrate them into AI generated content and here is a beautiful example of that and this is a brand new ad from original source and you can see here this is a real footage where we have the product being shown the shower gel and then suddenly we have a transition where the video changes from real footage into Ai and now we're floating down a slide with a million lemons into the ocean riding a giant lime and it's making the best use of AI here because it's featuring very fast-paced and challenging Transitions and then what we have is this beautiful transition from AI back to reality and we have our character Landing back in his mundane boring life and you can certainly see the Brilliance of this ad in how it is demonstrating that we can escape reality with AI and also by buying a zesty lemon and tea tree shower gel now I believe that AI is giving us huge opportunities to be making our own projects and to be profiting from those and because of this I've put together an ebook with some of the most exciting AI Side huttle Opportunities that you can try out I've put a link to that for free in the description below and I also would love to share with you this video which has a whole host of other exciting tools that you might not have seen before that I think are worth taking a look at I want to thank you for watching watching until the end end and most of all I want to wish you a delightful day
2025-03-12 17:54
