I’m building a simpler way to turn still images into AI videos

Hi everyone,

I’m building Image to Video AI as a solo side project.

I started working on it because the current image-to-video workflow often feels more complicated than it needs to be. Users have to compare different AI models, understand unfamiliar settings, estimate generation costs, and sometimes move between several tools before getting a usable video.

I wanted to create a more focused workflow:

1. Upload a still image

2. Describe the subject or camera motion

3. Choose the video model and output settings

4. See the estimated credit cost before generating

5. Preview and download the result in one place

The product is designed for portraits, product images, travel photos, property images, artwork, and other still visuals that people want to turn into short videos.

The first public version now includes image-to-video generation, adjustable aspect ratio, resolution, duration and frame rate, generation history, and several AI image and video models.

The hardest part so far has not been building the generation interface. It has been deciding how much model complexity to expose.

Advanced users want control over the model and settings, while newer users usually just want to upload an image and get a good result without understanding the underlying technology.

I’m currently trying to find the right balance between those two experiences.

I would especially appreciate feedback on these questions:

- Should the default experience hide the model selection completely?

- Which matters more for first-time users: lower cost, faster generation, or better motion quality?

- Does the homepage make it clear what the product does within the first few seconds?

I’m still early and working on the product, onboarding, and positioning. Honest feedback is welcome, especially from anyone who has built an AI generation product or a credit-based SaaS.