i'm not turning back whatever's inside I'll face it i know you're watching show yourself holy sh I'll show you how I've refined Google's VO3 AI video model so well that it cost me about $2 per minute of output it costs most people right now about $6 to $8 per second and that's quite expensive i'll show you how I was able to be effective with my outputs so it's not so expensive as AI right now as it's new it's still pretty pricey so I'll show you how to get that and exactly what you want more often and I'll show you how you can access this model with Google VO3 due to the fact that it might not be available outside of America since Google Flow Google Labs and Google Gemini are video creation AI tools that are only available in the states right now so let me show you how to break it down and the options that you might have even if you're not in the States all right so this is going to be a very detailed video on how to use Google Vo 3 now this is going to be with multiple different platforms to give you some options first there's Google Labs so let's go over what Google Labs is if you type into Google Flow F L you'll find that Google Labs is really what you want to use and it's called Flow or Google Labs it's the same thing so with Google Labs what you can do is under labs.google you can choose to access a project now when you want to put in your prompt be sure to choose the correct model here and I'll go over what these do later but ultimately this is where you add in your information from what I have experienced text to video is going to be the best option text to video is the most powerful and common use that people are using right now online frame to video is not really that powerful and ingredients to video is part of the Google ultra plan so you have to have access to that and even then it's not that good i've experimented with it and maybe I haven't been doing it right but it's still not as perfect as I want it to be it's more volatile than just text to video so that's Google Labs google Gemini is bas it's it's part of Google's product it's their same service but it's just on a different platform their Gemini AI service now with the Ultra plan you get access to five video generations per day with the lower plan I think it's about $25 per month or 20 per month if I remember correctly you get three video generations per day now you don't necessarily run out of credits you just have a maximum amount of videos you can generate per day and so when you want to generate something be sure to hit the video tab here or the video icon put in your prompt and you can hit enter and it creates your video for you now those two services Labs and Google Gemini you know both owned by Google they are only available in the United States at least right now this could change in the future but for now people can't access it and to be honest it is one of the more cheaper models out there for AI video generation specifically with VO3 which is one of the best video models out there we'll dive into a little bit of Clling 2.0 and things like that that will show you some pretty close competitors uh and relatively cheap considering but Labs and Google are available in the United States so let me show you a couple that are available worldwide so first it's foul foul.ai and they have the VO3 um essentially a plugin here so what you can do is put in your prompt like you would with labs or or Gemini uh and you can choose a couple parameters but really there's not a whole lot of um options except 9x6 is added here so if you want to do social media content this might actually work pretty well now there's not a lot of other things that would change your output that you can change here so really it's just putting in your in your prompt and then you can continue creating your your content now keep in mind that $3.7 is going to be uh charged for 5 seconds of video so it's about $6 for 8 seconds of generation which is quite expensive unfortunately that's just what is available to the rest of the world right now if you're using the Gemini app or if you're using Flow or technically labs.google it's a hell of a lot cheaper and I've been able to get it down to $2 per 60 seconds of video whereas on FAL it's $6 for eight seconds kind of pricey uh but you if you got the budget go for it now that's foul.ai there is no way to change um how expensive that gets with Google VO3 but there are a couple other alternatives so Paulo I'll leave a link in the description with my link it is an affiliate link you do get some free credits with it too so you can mess around with it first and give it a try i highly suggest that but with Polo what you can do there's a lot of other really useful tools but let's take a look at specifically the text to video option here there's so many options you can generate images which I've actually done um so if I just scroll up here you can see these images I've generated it's very cheap this is almost an all-in-one platform which is quite nice because then I can get my images generated of the character and then I can actually make the character do movement based on their photo so I can use photo to video generation and keep my character consistent which is one of the biggest issues with AI right now at least Google's VO3 so I'll show you how to do that later on in this video cuz this is kind of a all-encompassing video but ultimately textto video in Polo.ai AI is going to be a really really powerful tool and is worldwide again now it is slightly more expensive than Google V3 in fact sometimes it could be significantly more expensive so I kind of broke down some of the costs here so for every 800 credits you're spending about $30 if you are spending the I think it's $30 per month subscription now if you go up to 10,000 credits per month that's about $220 or $230 per month which you know Google Gemini/Google Labs Flow is $250 per month so you're still getting a pretty good deal out of that you're actually getting more credits for spending more so if you're spending round about the same amount for Google Gemini and Google Flow on Polo then you're you're essentially getting a a somewhat equivalent deal in a sense now if you're doing Google V3 cuz they have a Google V3 model right over here where it can generate audio with your video and I'll show you some of the results that you get with all these models uh then you're and if assuming you've spent 10,000 or or or acquired 10,000 credits it's going to cost you about 330 credits per video generation with VO3 which comes to about $7.2 per 8 seconds a little bit more expensive than foul.ai however you get a little bit more power and and flexibility when it comes to using this model in in in Apollo so number one you can generate your prompts here but if you wanted to do consistent character video what you can do is upload your images and choose 1080p let's say you want to do 10 seconds um you can put in your prompt with the images so I've actually used my image and it's it's done okay um it doesn't do very well with screenshots and that's what I used and so I actually generated images and used this tool and it actually did pretty well so I'll kind of run through that in some future videos and how to optimize this however um these are kind of the the price breakdowns here so VO2 which is 300 credits which if you're going to be doing VO2 you might as well do VO3 because it's extra 30 credits and you get audio with it so it's really not that big of a difference in price as you can see here $7.2 per 8 seconds on VO3 and VO2 is $6.58 per 8 seconds so you'll get an extra 30 uh an extra three videos with VO2 it's not really that much to be considered so um that's something to consider now cling is also a model you can use in here so in Polo you can choose Clling 1.6 2.0 there's lots of models that you can choose from and they all vary in price now one of the better ones in Polo is Cling 2.0 cost 100 credits if you get 10,000 credits kind of similar price to what you do with Google Labs uh it'll cost you about $2.19 for 5 seconds of video so actually not too bad considering all the other models out there and Cling 2.0 is not a bad model to consider so let me show you what that looks like so here's what the video generation with Cling looks like and yes it is doing an image to video and you can see the image that was over here it's similar to what the first frame is on this video and so it's still got to consider what the first frame is on of of a video from a photo sometimes it likes to take the as you see the photo here from as the first frame and turn it into the video you could always cut the first half second of the video if you really wanted to uh however this doesn't generate audio so if I go to cling 2.0 it doesn't generate any audio uh but if I do image to video you can also choose cling so there's there's multiple options that's 1.6 six cling 2.0 it doesn't generate any audio however if you want to do let's say Google VO3 it does generate audio and you can choose from a video image now I did that with a couple so let's take a look here so here's Google Vo3 with uh the image to video option on tight and cleaning hard so definitely a lot better and it's it's obviously got audio and it's got him talking so that's great that's with Google V3 again that's 330 credits which comes down to about $2.7 for 8 seconds so that shot cost $2.7 little pricey right definitely a huge difference compared to $2 for 60 seconds of video when I've used Google Gemini and Google Labs and I'll show you more about how to actually do that later on in the video so there's options here you can use Polo 1.6 six so you can get 400 videos with that with that model and Apollo 1.6 is actually not too bad so um I did some video or text to to video it doesn't do very well with that so I would just do text to video or those were image to videos i'll do text to video you can do use Polo 1.6 this is where you get 400 video outputs with the 10,000 credit subscription per month and you can see those prices over here when you click add more and you can choose what those prices are so you know around you know $220 per month google Flow is $250 per month um you can kind of see the the differences here and you definitely get a discount the more you spend per month especially if you're spending that much with um Polo.ai um Polo 1.6 this is the generation that you got and it doesn't do any audio but for for video I mean it's quite fantastic you can get 400 generations of this you can add in your own sound effects and stuff later a very cheap model there so that has actually come down to 55 cents for 5 seconds so pretty decent and then there's Cling so let's talk about Cling personally I like Cling a lot just because it is slightly cheaper than Paulo however there is a little bit more limitations when it comes to your outputs so if you want to use Cling 2.1 you can't do it on text to video which is quite frustrating at least for now uh it also doesn't do audio for text to video unless you choose I think 2.1 master or actually I don't even think it does it at all yeah and you can't even do text to video 2.1 which is the audio so if you want to do 2.1 you've got to do uh frames so this is almost like a a photo like what I've done here with the this photo that's extremely loud so even the audio isn't that great it's just scratchy audio so you could tell it not to generate any audio and then you've got multiple elements so you can add different photos and you can swap things and add it's just it's really powerful it's just too too deep for for today's video however with your photo you can add in your prompt and I did get this result it looks really really good considering all the other models and it did you know it's obviously not perfect however there are other options here that you can choose from so 2.1 master this is text to video without any audio and it doesn't generate any audio either and then there's Cling 1.6 so these are all you know further down we go the cheaper these models get so you can definitely see a difference in quality but maybe there's some videos where it doesn't require some crazy um high quality result and then this is Cling 1.2 you know it also looks quite professional but there's less of the cinematic feel and it's less prompt adherence um but then there's also the reference video that you can use it's included some sounds this is actually with the sound generator with with the button over here so quite an indetailed um and in-depth program here it's slightly cheaper than Polo however there's less things you can do with it so if you do want the cling video model you can actually still use it in Polo so that's that's an option there too so with all of these models sometimes it can get expensive especially if you're not inside the United States you don't have access to Labs and Gemini at least right now so here's how you can actually get access to it and you can get it for free for a month so you can just take advantage of that option and honestly you can probably create as many Gmail accounts as you want obviously you've got to put in your payment method there when you start the trial but you could take that out afterwards but let me show you how to do this all right so when you search flow on Google search engine this will come up and it's technically labs.google then you will be greeted with this screen now when you want to create a new project it's probably not going to let you generate anything because you won't have any credits so you need to subscribe so let's click subscribe and here you can get one month for free and subsequent months are $20 per month so get AI for free add in your payment information and then you'll be good to go for a month for free now if you don't have access to Flow and Google Labs and uh Gemini from anywhere else in the world because it's only available in America you can try and do this with a VPN now if it doesn't work with your current Gmail account you might have to create a new one with a VPN from inside the states and then once you've done that you should be able to upgrade to Google Gemini uh get access to that get access to the labs you'll get a,000 credits to use each month in Google Labs with the $20 a month subscription and you can just repeat that process for as long as you really want to and have as many Gmails accounts as you want and essentially have access to that for as long as you really want to now it's important to understand the difference between all the models when you're using labs.google in here when you hit these buttons you can choose between fast fast text to audio which is a beta and then VO2 quality which to be honest is obsolete at this point and then VO3 quality with audio i personally almost never use VO2 fast and VO2 quality in fact most of the time I'm using VO3 fast text to video because this costs 20 credits as opposed to 100 credits which is the higher quality audio version and honestly the VO2 still gets you know about 90% of the way there when it comes to comparison with the quality version of this and it still is able to output hyper realistic result and even if it doesn't get it on the first try you you can hit it again and it'll essentially use 20 more credits you get five chances to do that to hit the equivalent of what you would spend on VO3 quality by the time you've done it five times you've at least got one really really usable output and is very comparable to the VO3 quality so that's how I've been able to optimize this and so with this model I've actually been able to get my cost down to $2 per 60 seconds of video footage because of the prompts that I've also put into here so it really reduces the amount of mistakes that AI makes and helps it get it right more often so I'm going to show you how to do that as well so V3 fast at least right now whatever model is available in the future I'm not quite sure what that'll look like but for now this is going to be the best and cheapest model to use in labs.google now something to remember when you have that selected obviously choose one output because more than one output is just a waste of your credits what you have to keep in mind is if you go back to your flow tab here so the main homepage with your projects and then you go back into here it's going to reset the model so it's going to do VO2 fast without audio and if you want to do fast with audio you got to make sure to change that people have gotten that mistake many times where they think it's just stuck on there once they select it and that's good to go when really it's not the other thing you got to consider is if you refresh the page it's going to reset the video model as well so people have done that and have trying to figure out why their video hasn't been generating audio all right so now that we've gone over you know the different models that you have available in Google Labs what you want to also know is how do you get what you want you know get the specific results and uh consistency so I'm going to show you how to do that so let's go ahead and actually create something and what I've done is I've created two custom GPTs that will help you get exactly what you want 98% of the time it's obviously not a foolproof system but it's been what I've used in order to generate so many different outputs consistently and a lot of people have been asking me how I've been able to do this and this is part of the reason how I've been able to get this to be so cheap and the custom GPTs I will link in the description below go ahead and sign up make sure that when you sign up you've got to hit the confirmation subscription in your email and then you'll get your email with the GPTs that you can get access to remember to check your spam if you don't see it within you know about an hour or two but once you've signed up to get access to these custom GPTs you'll have the ability to create hyper realistic results and get consistent outputs in high quality so there's a version where it's just you know cinematic outputs and then there's another version where it's selfie style so you've probably seen those Yeti or Bigfoot selfie videos on social media this is the one you want to use if you want to be able to get the same results if not better so I'm going to just stick with the cinematic one for now just because I haven't been h uh I haven't been showing or or showcasing that on the channel as much as the selfie one so what I'm going to do you can do you can go two approaches here so number one is you can plan the video as you go so let's just say we want a let's say we want a sci-fi drifter with a man driving it in the sky through a cloud city at night with neon lights so it's not very specific there's I'm not really telling AI what the story is or what I want him to say uh but it's going to essentially come up with a random story and let's say you know it's describing the scene here it's being very specific so ultimately you want to make sure hyper realistic is in the beginning because you want to make sure AI knows that this needs to be realistic it's also specifying the type of shots what it looks like the type of person or man that is in here so he's unshaven um he's 30 years old things like that and then it's even told him or or told AI what to say he mutters under his breath "Every coroner is a new deal or new danger." Okay shot in cinematic style it's telling it that's shot in cinematic style what you want to do is let's say we don't want him to say this let's say we want him to say "I love the wind in my hair let's go faster." I would also like to have the cockpit open so that the wind is flowing through his hair so make sure to specify that in the prompt okay so now that we're kind of uh refining this output and kind of what we want cuz I'm I'm being very vague on this video on purpose because I want to show you how actually accurate this can get even with somewhat inaccurate prompts these are kind of just um on the fly and so what I'm going to do is I'm going to paste that in here make sure I'm using the VO3 fast text video with the audio option i'm going to send that now that's one way to do it and you can continue to adjust this as we go and continue to add to the story but what I'm going to do is actually let's say that I'm I have a story and a shot list planned already i would like to create a story with a beautiful dark hair pale skin woman in a red dress and hoodie cloak with a cape in a dark muted gloomy mysterious forest period i would like for you to sketch up a basic story line and also a basic shot list so the basic idea here is um a lady in a red dress in a gloomy forest and I want AI to kind of draft up a basic story and shot list so we're going to go ahead and almost refine this strategy maybe you've even got the story and you can just paste it in here and say come up with a shot list of what's going to happen and what it's doing here is establishing the different shots that are going to happen here so now what I'll do is uh say start with the first prompt and then it's going to go ahead and create that prompt for me so hyperistic wide establishing shot and that's based on this first shot that it has over here in the shot list and it's going to go ahead and create a story around what we've spoken about and it's going to create an accurate prompt for us jumping back into flow here with our first prompt that we added in it's still going so it does take a little bit of time however you can have five of these generations going at a time in Google Labs whereas Google Gemini you can only have one going at a time keep that in mind the other thing to consider is if your output is failing because sometimes it'll say fail to generate or or something along those lines uh you can simply just hit this button over here reuse prompt and hit you know send it again and sometimes it'll actually manage to do that but sometimes it'll keep having that issue what I've done is actually gone to Google Gemini pasted that same prompt in with the video option here and it's been able to get that output so sometimes it feels like Google Gemini is slightly a little more stable than Google Flow not entirely sure why um but something to know in case your output keeps failing the other thing that if it's going to be failing in its generation it probably means you have some type of copyrighted material in there for example let's say you can't generate famous people like Elon Musk or the Pope if you describe the Pope or maybe some type of trademark item it may not always generate however I've been able to generate Darth Vader so it's really just a hit-and- miss type of opportunity there that you have okay so now it's finished generating let's take a look and see what it looks and sounds like i love the feel of the wind in my hair let's go faster cool that's actually really good now this is how I've been able to get my outputs and sometimes if it doesn't look realistic again I would just hit the reuse prompt and hit it again and it'll just go through and give me a very different result most of the time now speaking of different results how do you get consistent results well I'm going to show you how to do that but first I think it's important to understand how to say uh or how to get your characters to speak in different languages so if you want them to speak in Portuguese Spanish German whatever it is how do you do that well I'm going to show you how to do that right now on me tight and clean we breaching hard so when using my custom GPT let's actually you know what let's try the selfie style just to experiment here so what we'll do is I'm just going to click on one of the presets here now when you click on this uh obviously you're going to get whatever the heck the the preset is but if you want to put in your own information your character who they are a little bit of their demographics things like that then what you can also do is include in that prompt make sure that these characters speak in whatever language so if it's Spanish you want to say make sure these characters speak in Spanish right now what we've got is English so welcome honor guests to my night of visions all right so English obviously we don't want that so let's do make them or let's do make the make the characters speak in Spanish okay so once it gets that it's going to go ahead and create that prompt and the quotations or whatever the characters are going to be saying is going to be spoken in Spanish and it's going to be spelled that way too so you can see the entire prompt is English which I suggest because it gets you the best output with Google VO3 on their uh labs.google.com which is over here or on their Google Gemini app uh that's going to get you the best output and then what's being said in the quotations is in the language that you want it to be and it's spelled in that language and it's also specified that they are speaking in Spanish right so whatever that language is now you can copy and paste that and what I'm going to do is I'm just going to do so and we are going to make sure we use the correct model here so I'm going to do VO3 fast text to video beta that's 20 credits as opposed to the other one which used to cost well still cost 100 credits but you don't have to do that anymore you could do the cheaper one and you may get some slight variations in terms of result but it's a lot cheaper especially if you get five wrong on that then you've spent 100 credits as opposed to the more expensive model which will spend 100 credits in one go so you've got essentially five attempts to get the same type of result or five variations of that if you're going to spend the same as the more expensive model okay so it's finished generating it and based on the prompt now remember we did this with the selfie style prompt uh right over here so it's going to do it in selfie style which is awesome so let's take a look and see what it looks like no idea what she said there but cool worked out so how do you get consistent character outputs cuz this is something that people have really been struggling with well number one is Paulo is really great at getting consistent results so consistent character video option here you can choose images of this person you only need one but you can do multiple and I've been able to get consistent outputs for example with this video here that um Black Swapman looks the same as over here using very similar uh image to videos and and text to videos and and things like that now it also depends on what model you're using so in for example this one I used Polo's AI model it's not great so sometimes what you might want to do is do image to video and what I've done is I've used Google VO3 to do this um and it's it looks really great so you can choose 1.6 which is their cheaper model here i haven't given it a try yet but it's definitely something worth considering it only costs 10 credits so let's give that a try cool so I'm going to do 720p at 5 seconds i've entered in a very simple prompt and I'm going to hit create we'll see what it does with the most cheapest model there now with text to video though so if we go to text to video what you can do is have the exact same description of the person and get really consistent output so this is what the video looks like from Apollo's model and looks really really good so that's one of the cheapest models you can use out there with 55 about 55 cents per 8 seconds of video output now if you want to get more consistent character results without using Polo cuz maybe you feel like it's just not the website I want to use here's how I do it with Google Labs and Google Gemini so with my custom GPT what I've made it do is if I were to essentially so this is the first prompt it had about the um the sci sci-fi scene that we created i'm going to hit or or type in next and it's going to create the next scene in this or the next shot in this scene keep in mind it's going to redescribe the man's character the way he looks his output and just his features because what you want to make sure is that you're redescribing every single little item so that the output is exactly the same or as close as possible it's never going to be completely perfect but the goal is to get consistent output and that's how I've been able to get consistent outputs with my characters is redescribing the person every single prompt because between this video and the next one I create in here it has no context and if you click the add to scene feature uh and when you hit the plus icon here and you can do jump to or extend it doesn't do it very well especially when you're trying to have consistency in the look of the character it's not going to look that great so let's take a look at what Apollo's 1.6 six model looks like with the image to video option there's not going to be any audio however wow that looks amazing and if I use the same image to generate another thing in the story so let's go to image to video i'll use the same frame here it'll let you choose the aspect ratio you can do 16x9 and you might get a 16x9 output uh but I'm just going to do or you could choose 9x6 excuse me for social media and stuff but I'm going to do 16x9 and then what we'll do is we'll just say or what I could probably even do is generate with AI i'll just add in what I want here and click continue i even spelled woods wrong it should be woods okay so let's choose that one so let's do continue okay and I will hit generate and then we'll see what type of consistency it'll create really what I'm considering and concerned about is how she looks and what the environment looks like what she's doing i can change what she does based on my prompts with chat GPT uh right in here and so if you have the custom GPT that I've linked in the description you can also use these outputs to get more realistic storylines and also make sure you get consistent outputs which is obviously super important so this tool here image to video is going to be super powerful it's actually better than Google Labs text or or image to to video i don't really know why but uh it it shouldn't be that but it is um so at least you've got uh Polo.ai AI to really help with that and again the link in the description will get you some free credits you can kind of mess around with and uh give it a try so let's see what we have the character doing in here so cool she's getting onto the motorcycle and driving backwards okay i mean that's not a big deal cuz we didn't spend much for that and so nice thing is is let's say at the end of the previous scene here uh let's go all the way down to somewhere here you know we can even choose the last frame of this and move on from that so and and provide that in the next image to video generation so that's something you can consider as well however if you wanted to you can also do consistent character video so it doesn't consider the frame and and start the video from there it actually you can just add the image so I'll click okay you add the image of the person and let's go ahead and and start a prompt here so if we go to chat GPT have a fair woman with dark hair a red hoodie and a cape and a red dress walking down a pathway from a gloomy castle in the late evening with fog and mist and leafless trees and she suddenly falls down on the ground and mud covers her dress everywhere so we're going to create a story line around this and it's going to create the prompt for me once it's done that I'm just going to copy that and let's paste that into Apollo with the consistent character generation oh and it does say I have to have under 500 characters so what I'll do make sure all of the prompts following are under 500 characters and so now in this conversation with the GPT it's going to make sure that it has all of the um character limits um specified there and we'll just repaste that in here let's just do 720p i'm just going to save credits for the sake of the examples here and we'll go ahead and click create in the meantime while we wait for that let's do um let's do next part of the story and it's going to talk about it's going to redescribe her which honestly we don't need for this particular model because this is consistent character creation cuz it's using the same photo of her yeah she's soaked in mud and soaked and kicked in mud so this is really continuing from the story so I'll copy this and then I'll paste it in here and honestly the only concern I have is that this isn't generating audio and Chachi PT is making her say things so what I might say is don't make the character say anything this is a silent film so you could you know update that so let's actually copy that so that it may not um clash with the output here cuz there's no there's no audio settings here yet unfortunately um but that's where the text to video comes in which is quite nice um but I'll go ahead and and choose and it's going to go ahead and cue those up for me while I'm waiting for those if I go to image to video which I think is the more powerful feature here in Polo and I'll go to Google Vo3 I'm going to get the same prompt that I had earlier and one thing you want to see here is that there you won't have to worry about the character limit so I'm going to go ahead and go into here i'm going to You can choose an image too so actually I should be on text to video here and let's go to Google Vo3 i'll paste that in here with the prompt 1500 character limit i want to also make sure I'm generating audio it's going to cost me 330 credits which I have enough right now all right so I'll go ahead and hit create and then what I'll do is I'll go back to GPT and copy the second prompt that we had copy that and actually I'm going to make sure that chat GPT redescribes this because I told it to limit the characters to 500 what I'm going to do is start with this i'm going to start a new GPT just for the sake i want to show you what it looks like to go with um with the VO3 in Polo cool so we'll copy that and we'll go to Polo and then add that oh we need to go to this one instead because we've already generated this one so I'll paste the second one in the story here and hit create and so what I'll do is on the first two cheaper generations here with the consistent character it's finished the first one here so let's go ahead and play that so there she's in the mud and struggling that looks kind of funny uh but it's done really well considering it's a it's the cheap model so it's it's much nicer uh but we'll wait and see what happens with the second character generation here in the story but it's essentially had the person walking or it's actually the previous one here so it finished the mud one so I think the first one is taking a little bit longer okay so let's see what it did here cool she gets out of the mud and she's walking backwards for some reason but that's the thing with AI it's never perfect even with Google Flow's options here it is more consistent in terms of the output with characters but it's the mo it's the best model out there and it's just what is is the best right now until that changes who knows how that's going to change so let's see what Google V3 looks like so she's walking down here she didn't quite fall she just got down into the mud and then the next shot here is her standing up and getting out of the mud her character her her face looks very very similar to the first one um so that is done really well she's wearing this the same stuff um and I think this is the part that people have really struggled with is trying to get consistent character output i'm not turning back wow that's actually impressive i'm That's cool it got a really really great output there so see we've got thunder in the background it's just not turning back very impressive paulo great job i mean using VO3 and and getting some really great results there all in all VO3 is the model that you want to use it is the most expensive however if you're using Google Flow and unfortunately if if you're outside of the states you can't use it unless maybe you signed up with a Gmail account and you are using a VPN then maybe you could actually get access to it and even use the free trial so people have still been confused as to how I get really great results so I'm going to actually run you through how I do that so let's go ahead and choose this story so I'm going to choose the first shot here of the woman let's go to V3 i'll paste that in here and I'm going to do V3 fast text to video i'll hit enter right there and then I'm going to go to the second output here and I'm just recreating what we just saw in Polo kind of compare and see how it does what I'm going to do is hit next in the GPT so I can create the next part of the story okay and I can I've created in a way where you can just click copy at least this is what it should be doing most of the time and then I'll click copy and I'll just paste it in here it's using again VO3 fast text to video um remember you want to make sure that it's using the right model so I'm I'm generating some outputs here and I'm just going to hit next again and as it's creating I'm just continuing through the story now normally I'd be checking these and saying okay this is what I this is what I want to happen because I haven't really given it a a guide here as to what I'm doing with the story or where it's going so normally I'd be checking what it's doing and what it's saying but I'm just doing this for speed and simplicity um and just to show you what these outputs all look like but you'll notice that it's describing her so with long hair wavy dark hair her face pale and marked with streaks of dried mud so it's considering the past story here too which is nice that's really the power of chat GPT and the custom GPTs here and so it's redescribing the person each time in these prompts because it doesn't have the context of this image if you are doing this um with the image the charact consistent character video here you don't have to redescribe her each time you can just describe what's happening next um image to video you want to be careful with this because it's going to use that image for the same output uh it's going to have that image in the output of the video and then morph into the rest of the video consistent character it's going to essentially have it as an element and it's not going to use that image as the first frame and then text to video the sky is the limit and that's kind of a similar way that I'm using here so let's see how it's created this using the fast version of VO3 not bad okay uh now obviously there's no context between these shots so there's obviously going to be some slight differences but let's take a look i'm not turning back nice she does look very similar there's obviously not going to be 100% shot for shot or or just person to person it's not going to take the person from this prompt to this prompt but the point is the the GPT you can see is describing her exactly the same each time so that AI is you know creating the person 95 to 98% of the way there each prompt sometimes it'll go way off and that's when you re regenerate it and that's the beauty of using the freer or not freer the the cheaper version of the VO3 so you can essentially get five attempts as before when it used to just cost 100 credits and you didn't have any other options it wasn't cheaper let's see what the next shot is whatever's inside I'll face it that's impressive i know you're watching wow pretty cool so if this video was helpful I do have some other videos on using Google Vo3 and especially for the selfie style you know when the you know yetis and Bigfoots are holding the camera like this in selfie style talking to the camera i'll leave a link to that video over here on how to do all of that but hopefully this video was helpful and provided a lot of insight as to how to maximize VO3 and how to make it cheaper and just everything that's optional and and involved with it so click on the video here before it disappears hope you enjoy it and I'll see you