No comments yet.
Been a full-stack dev for a while and got deep into the generative AI image rabbit hole. Wanted to share my newest project, BestPhotoAI.
It’s an AI photo studio that lets you train a model on your face and then generate hundreds of pictures of yourself in different styles headshots, Tinder profiles, that kind of stuff. There's also a bunch of free single-shot(vibe coded mostly in last 2 days) tools (over 40 of them) for things like changing hair color, removing objects, or turning yourself into an anime character and animating images into videos.
This whole thing started because I was running another AI site with just 4x 4070tis which grew to 5k/mo rev.(AIEasyPic) but then I got a bit tired of my current tech stack, the bad SEO I got by scraping and reposting civitai models and the fact that if you are using 2 diff. lang. for backend and frontend cursor feels like a drag. So I did the only logical thing and just vibecoded a clone of my own product in less than a week before going on a holiday for 2 months which was probably a bad idea.
The tech stack is simpler now, basically all Next.js(with trpc) so cursor can basically one-shot an entire page with functionality in a single prompt. Making the pipeline itself on runpod was the hardest part, but thats basically a solved problem with my exp. with previous AI images product. Outside of HeadshotPro which has probably if-elsed out a lot of the edge cases(eye color changes, niche "races" getting whiter or darker, etc probably solved by VLM-ing the images first), I can say it truly is the BEST site for training a model of yourself and trying out a lot of the newest AI image/video models that are released.
Also I figured the HN crowd might get a kick out of the free stuff. No signup, no BS, just upload a photo and mess around with like 40+ tools I've vibecoded in the last 2 days:
You can try the free tools here: https://bestphoto.ai/tools/all (make sure to select the ones that say "Free-to-try")
Happy to answer any questions about the tech, the struggles of an EU-based AI startup, or difficulties with running a (mostly) self-hosted AI startup or ask me if I regret starting and getting way too into AI image gen side instead of making an AI text product(yes I do and a lot). Any feedback is gold.
Let me know what you think.
