After months of building, testing, and improving, I finally launched UluP Spaces , a visual collaborative workspace where projects become interactive maps.
It's a small tool that solves one specific friction: when AI generates an HTML prototype in minutes, but getting feedback on it still means screenshots, red circles, and pasting into chat apps.
This will be my first launch on Product Hunt. Wish me luck !
I m building a product that is already helping me (during testing) organize my lists of favorite places (cafe bars, restaurants, bakery, etc.) in different cities. For example, I live in Spain, in Valencia, and it s actually hard to find really good coffee in caf s here but I managed to do it. I created a list of great coffee places in Flistoo and want to share it with people who are looking for spots like these. And anyone can create a list of their favorite places and share it with friends.
Flistoo helps you discover, save, and share food and drink spots around the world, so you can stop saving places in notes and easily manage them in one app.
Couple of days back, I posted about using svg.comicbanana.com to create anonymous profile pictures. A good samaritan on the internet suggested exposing this functionality, so now you can use this API to configure and fetch comic like SVGs.
API doc with samples: https://svg.comicbanana.com/?men...
API to fetch random SVG avatars: https://comicbanana.com/api/svg-...
OrienFrame is an AI portrait studio inspired by classical East Asian visual aesthetics. Upload a portrait, choose a curated style, and create an artwork with a quieter and more intentional visual direction than a typical collection of AI filters. I built OrienFrame because I felt that many AI portrait products offer a huge number of filters, but very little consistent art direction.
OrienFrame takes the opposite approach. It focuses on a curated visual identity inspired by East Asian aesthetics and presents the experience as a digital portrait studio rather than a general-purpose image generator.
The product currently includes one free generation per day after Google sign-in.
I d especially appreciate feedback on three things:
Last year, our data pipeline kept failing in ways that made no sense. Requests would work perfectly for hours, then suddenly start returning 403s and CAPTCHAs. We blamed everything our code, our infrastructure, even the time of day. Six months of debugging later, here's what we actually learned.
1. IP freshness matters more than pool size
We assumed a large IP pool meant we were safe. Wrong. We discovered that many providers recycle IPs aggressively the same IPs get reused across customers, and once an IP gets flagged, it stays flagged. A smaller pool of freshly validated IPs consistently outperformed a massive pool of stale ones.
As an indie developer, I noticed how exhausting it is for cross-border and vintage sellers to manually draft titles, SEO tags, and material attributes for every single unique item.
So I built VintageLister AI as a quick MVP to solve this exact pain point!
