I purchased Headshotly.ai to create YouTube thumbnails. I realized that I had missed the chance to buy PhotoGPT, but then Headshotly.ai came along and matched exactly what I was looking for, so I immediately bought the Tier 3 plan that supports YouTube thumbnails.

Many users advised taking the initial model training very seriously, so I spent about three days selecting 12 photos of myself, taken indoors and outdoors, in different weather conditions, and wearing different clothes. However, unlike the guideline on the website that says 8 to 12 photos are ideal, the system currently only allows a maximum of 6 photos to be uploaded. These six photos are subject to very strict requirements, and any photo that does not meet the criteria is rejected. I now understand that this is designed to ensure successful model training even for beginners using the minimum number of images.

The quality of the thumbnails I generated is truly impressive. They really look like me. My biggest concern was the text on the thumbnails. Since I am Japanese, I want the thumbnail text to be in Japanese. When using “Clone YouTube,” I can input a reference video URL and it generates cloned thumbnails with accurate Japanese text in several variations, which is excellent.

What I would like to see improved is the “Custom Script” feature. Even though I wrote the description in Japanese, the generated thumbnail text was in English. If there is a way to generate Japanese text on thumbnails, I would appreciate it if you could let me know. If not, I strongly hope you will add a feature that allows users to specify the language of the thumbnail text.

Other than that, the service is perfect so far.