サービス提供：Pollo.ai
A lot of different generation AIs have been released in the past two years, but they are all very high quality, so it's difficult to say which one is the best.
I'm trying my best to keep up, but it's quite difficult to cover everything.
Meanwhile, I got a slightly unusual video generation AI called Pollo AI, so I would like to try it out. There are some features that I've never tried before, but it's fun. But I'll be careful and give you an honest review.
- New to synthetic video?
- People who want a simple interface
- Those who want to complete all video generation AI functions with just one tool
- People who want to make things easily without special skills
- People who can wait 4-5 minutes for the video to be generated
Overview of Pollo AI
Pollo AI is a next-generation video generation tool that uses AI technology to generate high-quality HD videos. Its biggest feature is that it can automatically generate videos in a wide variety of styles just by inputting text prompts, images, and videos. It is especially suitable for producing professional-quality videos and advertising videos.
Pollo AI is equipped with multiple AI models, and models such as Kling specialize in realistic video expression, including Japanese people. It also has an "upscaling function" that converts low-quality video into high-resolution, detailed video, making existing materials even more attractive. The intuitive interface allows you to create professional-level videos in a short time, even without specialized knowledge.
In addition, with its easy-to-understand interface and simple operability, this is a video creation app that can be used by beginners and professionals alike.
In addition, Pollo AI is an efficient solution that reduces costs and shortens production time, opening up new possibilities in video production.
The official Pollo.ai website is hereEt al.
As mentioned above, the biggest feature of this app is that you can choose from multiple well-known AI models to generate videos. Each has its own strengths, and the following AI models are available:
- Chicken1.5
Pollo 1.5 is an advanced AI tool that can generate high-quality HD videos from text and images. It supports resolutions up to 4K and features realistic and consistent visual expression. Videos are basically 5 or 10 seconds long and have an intuitive interface that is easy to use even for beginners.
- Kling 1.6
Kling 1.6, powered by Pollo AI, improves image-to-video generation by 195% compared to the previous version, with enhanced prompt comprehension, motion dynamics, and visual quality. This enables higher quality and more consistent video generation, setting a new standard for AI video production. It can generate 5-10 second videos.
- Kling 1.0
Kling1.0 is a model that can generate videos at a lower cost than Kling1.6. If you don't want to spend too many credits, you should use this model.
- Vidu 1.5
Vidu 1.5 adds a new feature called "Multiple-Entity Consistency" that allows multiple characters and objects to be represented with consistency, making it possible to generate high-quality videos even from a small number of reference images. In addition, the generation speed has been improved, and a four-second video can be generated in about one minute. In addition, it supports a variety of styles, including anime and realistic.
- Hailuo AI
Hailuo AI is an AI tool that automatically generates high-quality videos from text and images. It is particularly good at reproducing detailed emotional expressions and actions, such as facial expressions and mouth movements. Even beginners can easily operate it, and it can be used for a variety of purposes, such as short videos for SNS and business presentations.
- Luma AI
Luma AI is an AI tool that automatically generates high-quality 3D models and videos from text and images. In particular, it is equipped with the latest model "Ray2" that realizes natural and smooth movements by understanding the laws of physics, making it possible to express complex scenes and characters. Its intuitive operability makes it suitable for a wide range of users from beginners to professionals.
- PixVerse V2
An advanced AI video generation platform released on July 2024, 7, that can create high-quality videos up to 24 seconds long from text and images. Compared to V8, the quality has been improved, enabling consistent video generation across multiple clips. It also supports a variety of styles, including realistic, 1D animation, and anime styles.
- PixVerse V3
An advanced AI video generation model built into Pollo.AI. It has improved understanding of complex user input and a new lip sync function that synchronizes voice and lip movements. Users can generate videos in a variety of styles, including anime, realistic, clay, and 3D, and can also use special effects such as zombie mode and monster invasion. It also supports multiple aspect ratios and is suitable for use on a variety of platforms.
As you can see, it is equipped with a number of well-known AI models.
On Pollo AI, different AI models have different interfaces, generated output quality, rendering times, and credit consumption.
Try out different AI models and find your own video generation AI.
Pollo AI User Interface
This is the default Pollo 1.5 screen. If you use other AI models, the layout will be slightly different.Prompt fields, aspect ratio, public visibility, copy protectionYou can operate the items such as. Enter the actual command in the prompt field. As I will explain later, the prompt field contains "Created with AI" option, which will enhance the prompts with generative AI.
And "Public Visibility" is an option to set the visibility of the generated video. This setting allows you to choose whether to make the video public or share it with a limited audience. For specific settings and details on the options, please check the settings screen of each tool on the official website.
And "Copy Protection" However, you can set it so that the video you generate cannot be copied by other users.
It has a very simple interface which I really like.
The right column shows the videos you have created so far.
I actually tried using Pollo AI
I basically compared it against the default "Pollo1.5", "Kling1.6", which specializes in high-quality depiction of people and other subjects, and OpenAI's "Sora" for comparison.
First, let’s give Pollo1.5 the following prompt:
As snow falls steadily across the city, a girl is surrounded by illuminations and looks up at the night sky. The camera slowly circles around her from a low angle.
So the video below was the result.
It's a pretty good atmosphere, but it's not snowing. (Maybe it's just falling a little?) And the camera isn't moving. It's a shame that the prompts can't keep up with the quality of the image.
And if you look closely, you can see distortions in buildings and other structures. It would be a bit difficult to use this for work.
Let’s try enhancing the prompt with “AI-generated” as an example.
Click "AI-Generated" in the bottom left of the prompt section.
Prompts are reinforced in English.
Prompts from generative AI like this may produce better results in English than in Japanese.
The resulting video is shown above.
It doesn't look like much. There's no snow, and the camera work is elevated, which is contrary to the intention. Is it a problem with Pollo.AI, or is the prompt bad?
I would like to try it out using Kling's AI model.
With Kling 1.6, the number of items that can be configured in the interface has increased.
Immediate effect
Maybe it's a matter of following my prompts exactly, or letting the AI be creative. I'll leave it somewhere in between, which is the default.
Video length
Pollo 1.5 can only create 5-second videos, but Kling 1.5 can create 10-second videos.
This is a big advantage. *The credit consumption will also increase from 10 to 20.
Negative prompts
As the name suggests, this allows you to specify elements that you do not want to be added to the generation. It is not always reflected, but it can have some effect.
And below is a video created using the same prompt.
At first glance, I thought, "This is it!" I made a video that met my request. However, the snow is a bit questionable. There are some places where it's falling and some places where it's not. That's the only thing I'm disappointed about. The camera work is perfect. It circles around the girl smoothly. The buildings in the distance are not damaged either.
feel well.
Next, we’ll try generating a video of OpenAI’s Sora using the same prompt.
Sora just doesn't feel unnatural.
It follows prompts well and meets expectations at a high level.
It naturally realizes the "surrounded by lots of illumination" part that was missing in other models. It's like a step up from before... OpenAI often does silent updates, so it's a quality that makes you think the model may have been upgraded.
We measured the rendering speed for each model.
- Pollo 1.5…6 minutes 8 seconds
- Kling1.6…3 minutes 33 seconds
- Sora...1 minute 56 seconds
The results are as above.
Sora renders the entire length in just 10 seconds, which is incredibly fast.
If you want to do creative work with Pollo.AI, Kling 1.6 is probably the place to go.
My honest impression of Pollo 1.5 is that the waiting time is a bit long.
Other features
Other features are summarized in the left sidebar.
What's great about this app is that it has a full range of multi-modal functions as a generator.
These are all very appealing features, so I'll evaluate them from top to bottom.
Image to video
First, I tried out my favorite feature, "Image to Video."
The AI model used is Pollo1.5.
The image I used was based on a still image of the author eating a Korean thing that was popular for a while.
And the rendering result is:
This is it!
He's eating it like a pig! And it's of amazing quality!
I think this quality is equal to or better than OpenAI's Sora. Furthermore, Sora is strict about portrait rights and will give you an error when you try to render an image of a person.
But this doesn't give any errors. Also, I should mention that I made a mistake with the prompt and generated this video from scratch, but it seemed to recognize the scene properly and created a kissing video on its own!
It's normal for this kind of video to have fingers being torn off or regenerated, but this video had some suspicious parts, but it's within the acceptable range. It can be used as an alibi video.
It was made to be 10 seconds long, but the rendering took only 3 minutes and 57 seconds.
Well, Pollo 1.5 can be done if you try.
This makes me want to try out different things.
Consistent character videos
I believe this is a feature that is unique to this app and not available from other companies. What does it do? It allows the user to set a specific main character, and then create a scene by placing items and spaces around that character.
Let's actually give it a try.
This function does not include a drop-down list for selecting an AI model, so it is unclear which AI model is being used. There are three fields for inserting images, so I prepared three: the main character, the supporting coffee cup, and my two favorite Darts Live cards. In other words, these three items are the main axis and compose one scene.
I chose 8 seconds as the duration and 720p as the resolution.
It says "operating range," but I think it probably means the magnitude of the movement. For now, let's call it small.
However, this feature costs 100 credits.
If you don't save...the result is...
The scene was generated using three elements. *The way the cards are handled is a bit strange though.
This may need some more research.
It's especially strange that there are three cups of coffee. I tried a few times, but it was difficult to make it stable because there was only one coffee cup or one card. It's difficult.
Video to video
This is a feature that is also installed in Sora. It is a function that generates new videos from videos. I would like to give an honest review to see what kind of performance it will show.
By the way, I used a video generated by Pollo as the original video.
I tried rendering without any prompts, applying the Modern Anime v3 style. There are a variety of presets available in this style, and you can apply the style with one click and render.
The video above is the original video.
Wow! It's been converted into an anime style really nicely.
The idea behind this feature is that by taking live-action pictures of people and animating them with AI, it might be possible to create a coherent anime work. This feature can create videos up to 30 seconds long, which is long enough to make one cut, so I think it's quite feasible.
If you check the option "Change style of subject only," you can choose between "People" and "Custom." If you choose "People," the AI will determine this and apply the style to people only.
If you select "Custom," you can specify an object and change the style of just that object. This is an incredibly fun feature.
This is just speculation as we won't be using "Custom Background" this time, but it allows you to change the background of a character to a specified fill color.
Below is the result of changing the style to "Disney Pixar v3" and "changing the style of the subject only" to "People".
It's kind of scary to see him zooming in on a smiling face without changing his expression lol
This may be improved by prompting the patient to blink or change their facial expression.
The background is a real photo, but unfortunately there is a bit of flickering. If this could be dealt with better, it could create an interesting piece of work.
Make videos more beautiful with a variety of "AI tools"
Also"AI toolsYou can use it to edit various videos.
This is another professional feature not available in other video generation apps.
For example, a "video upscaler" improves the image quality of videos that have insufficient resolution by upscaling them. A long time ago, this was sold separately, so we're in an amazing era.
There are also various other tools available, such as repair tools and motion brushes, so it's fun to try them out.
Price Plan
Prices start from $10 per month for the most affordable Lite plan. You get 300 credits every month, so if you use it strategically, you can create 30 videos per month. The most expensive plan is only $25 per month, so it has all these features and is easy to use, so some people might want to use the Pro plan.
Summary: The future depicted by Pollo AI
My impression after actually using it was that it has a sophisticated interface. The UI design is excellent, and it's very easy to understand where things are and what their functions are.
Even though it has so many functions, you never get confused. You can understand it intuitively.
I felt that this was absolutely a good app for someone new to video to use as their first generative AI.
I wonder how the quality will fare when it goes head-to-head with Sora, but there are some outstanding aspects and with the expectation that there will be many more updates to come, I feel this is an exciting app with great potential for future growth.
As I look towards the future of Pollo AI, I feel motivated to continue to work hard.
![The magic of "fragrant"... How to use "solid perfume" in a chic and fun way for both men and women [Review]](https://mainoriti.com/wp-content/uploads/2025/02/IMG_0065-3-485x273.webp)
