Great results. How are you training the user updated images? Do you run on your GPUs or rely on third-party APIs? Lately, there are a lot of AI image-generation startups coming up. Thanks to Stable Diffusion + Dreambooth training which makes things possible.

Here are some more list

- Profile pictures/Avatar - https://PicasaAI.com

- Food - http://unrealmeal.ai/

- Hairstyle - https://www.hairgen.ai/ https://www.hairstyleai.com/

- Photo restoration models - https://replicate.com/tencentarc/gfpgan