Marius

✨ Photo AI just reached a new record of $150,000/mo 💳 2,573 active subscribers 💸 87% profit margin 👌 100% bootstrapped + $0 funding 😊 Employees: 1 = just me on my laptop 🥞 Tech: PHP + jQuery + SQLite on a Hetzner VPS with Nginx and Ubuntu It took 2 years, 7 months, 14 days, or almost 1000 days to get there Photo AI started out in 2022 as Avatar AI (the first app to make those gimmicky avatars which started that trend) Then I pivoted it into Photo AI a year later to take real AI photos of yourself. Because I thought that might be possible next. In a way I was way too early and the photos were actually pretty bad for about 2 years (remember the hands?). Faces were distorted and it didn't resemble you much. I used Stable Diffusion back then and it was the best model but it wasn't so good at people. But with a lot of help from @philz1337x I then built a giant pipeline of models on top of other models to fix faces, hands, and bodies and upscaling it to make it look somewhat good, but still it wasn't great In a way I had product-market-fit (PMF) but without the product. Because I knew the demand would be there but the product, well the technology, definitely wasn't there (yet!) Then in 2024, a new AI image model called Flux was launched and overnight Photo AI was now good enough: the photos actually really resembled you, the quality was high and hands, faces and bodies looked real I now did have PMF! Since then Photo AI has grown a lot and it's now become a highly realistic AI photo and video production studio for content that features real people. There's other AI image generators, but they're not as good with people as mine, because that's my focus and niche Photo AI is now used by fashion designers (to try on clothes), ad makers (for talking videos), marketers (to show people presenting physical products), education companies (to teach people courses), short film makers (making videos with real actors) and game designers (to create 3d characters from people) And even just people who want to do a photo shoot for their social media profile or dating app and can't afford an expensive $1000 photographer (you get 1000 pics for $49, so about 20x-50x cheaper) The app itself is standing on the shoulders of giants, that I am grateful every day for, which are the countless AI models I use and combine which are made by companies like Black Forest Labs, Bytedance, Kuaishou and more recently Google (with Nano Banana) made by very smart AI researchers and my GPU providers FAL and Wavespeed hosting them reliably When I started people kept telling me I just made a "GPT wrapper" (of course it didn't even use ChatGPT), then a "Flux wrapper". I now use a lot more models than Flux, so yes I guess it's more like "shitloads of models duct taped together with a lot of frontend and backend logic added wrapper" but that doesn't roll of the tongue so easily 😂 Most of my traffic comes from SEO and specifically pSEO, there's a lot of sub pages I added (like the photo packs and individual phtoos) which capture a lot of traffic and sign ups. X is about 9% of traffic and even less of paid signups, so a lot less than people usually think. I'm mostly on here to tell my story, X doesn't actually help so much revenue wise. Increasingly ChatGPT etc is becoming a bigger referrer of traffic and paid signups though There's so many other AI apps out there, and mostly we all use the same models, and I guess Photo AI shows there is enough pie for everyone, it's not the biggest either, but it's big for me and it's nice in its own way and I get happy working on it It's been a wild ride, and stressful at times (especially during 2023 trying to get Stable Diffusion to work), but mostly just really exciting to be able to work with groundbreaking technology every day and get paid for it too I remember in 2021 with my businesses Nomad List and Remote OK sitting and thinking "after almost a decade coding, everything is finally kind of stable now, now I am satisfied" but then the AI wave started in mid 2022 and I just knew I had to jump on it and catch that wave and I did with my entire being :D And I caught it I think! 🏄‍♂️ Surf's uppppppppp 🤠

🎉 Exciting News! To thank everyone for your enthusiasm, we’re launching a new activity! ✨ Post your creations made with Wan 2.5 ✨ Tag @WaveSpeedAI 💰 Get 10 credits for your submission 🌍 Open to all creators — join now and show your creativity! #WAN2_5 #WaveSpeedAI #AIVideo #CreativeChallenge #ContentCreation

Kling 2.5 is among the best video models at the moment. It very accurate when using start frames so you can easily stitch multiple videos together and make a continuous shot. It’s now available on @freepik