Lovable × Bloom: The AI Complaints Center

On the afternoon of Wednesday, October 7, we held the Bloom × Lovable Buildathon at PUBG Seongsu, a venue in Seoul's Seongsu neighborhood. The concept was an "AI Complaints Center." Starting a week before the event, we collected complaints through the Complaints Center channel on the Bloom Discord and the sign-up form: things people had always wanted to build with AI but never managed to. On the day, people took each other's complaints in person during small-group networking. Then every participant picked one complaint and built a product for it in Lovable in about an hour and a half. That's how 30 services were born.

A complaint means someone will use it

The event's theme, Taste Is All You Need, means that now that AI writes code quickly and easily, what matters more and more is the taste of the person building. So we designed the day around first hearing each other's problems and then solving them ourselves. We told people three lines were enough for a complaint: who would use it, how they handle it now, and what they wish existed. Our examples were a task that repeats every week that nobody has automated, a side project you've had in mind forever but never got past the first screen, and a club you're still running out of a spreadsheet. Junsu Kim, one of the Bloom organizers, explained why we bothered collecting complaints at all by pointing to customers.

If there's a complaint, that means there's a customer. As soon as you build it, someone who'll use it already exists.

Junsu's view was that even if the people who came don't write code themselves, they know their own problems and what they want to make. Hearing other people's problems also gives you a lot of ideas for what to build. If you sit alone with AI and think about it forever, the answers turn out more limited than you'd expect, so Junsu told everyone to ask as many questions and have as many conversations as they could. People living such happy, peaceful lives that they have no complaints could solve someone else's complaint and practice their building skills that way. As for why we cleared out a whole weekday afternoon, Junsu drew on personal experience.

You always tell yourself you'll work hard on the weekend, but when you try to set aside time alone on a weekend, human willpower is weak and a lot of the time you don't do it. I figured if we got everyone together on a weekday, you'd be able to focus much harder, so that's why we invited you.

For freelancers, college students, and job seekers, a weekday afternoon is the most flexible time they have. For people with jobs, we hoped it would be worth taking a day off for. From the start, we also wanted whatever people built to be usable right away in their own businesses or portfolios.

Networking with the recorder on

Groups were random, but only for networking. Building was an individual event. Networking ran 30 to 40 minutes: briefly introduce what you do, talk through problems you've been running into or what you want to build today, and then discuss how to build it. Junsu also suggested keeping a recorder running during the conversation. Talking it through organizes your thinking, and if you feed that recording straight to AI, building goes much faster.

A QR code above the networking area let people sign up for the Bloom app (app.bloomworld.ai). The app is invite-only, so Junsu used admin access to invite everyone who scanned it that day in one batch. When you sign up you get a profile member card, and you can connect with other members' cards. When a connection hosts a Mixer (our small curated meetups), you get a notification. The Bloom app's goal is to be an offline LinkedIn: the more you meet people, the more you keep the ones who share your wavelength and values. Another QR code led to the Discord Complaints Center channel, where we gathered all announcements and links to what people built.

Lovable: describe it and it becomes an app

Every participant used Lovable, an AI app builder that started in Stockholm, Sweden. You describe what you want in words and it builds a web app with screens and features, and you can publish the finished app right away at a lovable.app address. It passed $100 million in annual recurring revenue eight months after launch and is counted among the fastest-growing software companies. Its pitch is that people who don't write code can build products too.

Every participant got 100 Lovable credits and built from 2:00 to 3:30 p.m. Of the 30 projects, 29 were published on the spot at lovable.app addresses that anyone could open right away. An assistant that keeps your schedule in sync across any device via Google login, an AI that reads a single screenshot and tells you whether it's a scam, a service that pulls shelter notices from across the country through a public data OpenAPI, a room your whole group joins through one link, and a tool that turns a planning doc into presentation slides and a PDF all came out of roughly an hour and a half. We decided to collect links to the best projects and pass them on to Lovable. Nothing is confirmed, but just as people who made strong projects at our events with Genspark and Higgsfield got generous extra credits, there may be follow-up support here too.

Apps that take care of you

"Shwimpyo" (Comma) (dev-chameleon.lovable.app) is a wellness planner for freelancers working in Seoul. It finds the gaps between appointments, suggests places or ways to rest that fit those gaps, and tells you what to bring before your next appointment. The first screen states up front that it doesn't collect your current location and stores all data only on your device.

"DOERS" (start-your-doing.lovable.app) is an AI wellness app with the slogan STOP SCROLLING, START DOING. It tells you not to stop at watching and saving, but to actually do one thing today, and helps you keep up small daily actions for yourself.

"Star by Star" (sky-lit-tasks.lovable.app) is a casual to-do game. Each time you finish a task, a star lights up in tonight's sky, and when you finish them all a hidden constellation appears. The constellations you collect pile up month by month into your own universe.

"My Manager" (weather-routine-planner.lovable.app) is an AI assistant that organizes your day around the weather, your routines, and your schedule. Its maker called it a manager to crack the whip and keep their slacking self working hard. Log in with Google and you can keep managing your schedule and routines from any device.

"Today" (pixel-perfect-showcase-0870.lovable.app) is an AI Chief of Staff. You work with AI to spell out the life you want, it builds a path to that future, and then it decides what you should do today. In the example screen, your future self and today's schedule sit side by side, and the Chief of Staff suggests spending 1 hour 30 minutes of today's free time on your most important long-term goal. The app's tagline says it connects today to the person you want to become.

"The Grand Interfaith Prescription" (bible-comfort-card.lovable.app) has you choose what you're feeling right now (anxiety, exhaustion, career worries, loneliness, anger) and write what's on your mind in under 300 characters. It then prescribes the perspectives of three traditions, the Bible, Buddhism, and philosophy, bundled into one letter. It cites its sources too: Bible verses come from the Korean Revised Version, and the Buddhist and philosophical lines were adapted from classic scriptures and writings.

"Maieu" (maieu.lovable.app) takes its name from maieutike, Socrates' midwife method of questioning. It's a workshop for writing and thinking where Maieu asks the questions and you do the thinking, and AI doesn't write for you. Its maker described what prompted it:

As we've started using AI, everyone's productivity keeps going up, but along the way our ability to think and wrestle with things is shrinking, and the record of our own thoughts is getting fragmented and scattered.

Maieu offers to be a 10-minute sparring partner who finds the holes in your logic before anyone else does. It doesn't write for you. It only marks weak logic and gaps in red pen and leaves the rewriting to you. There's a capture mode for piling up ideas without questions, and a deep-dive mode where you answer questions, rewrite in your own words, and note the questions still unresolved. The app suggests using it to find blind spots in a proposal or deck before your boss or client does, and to rewrite a LinkedIn post or cover letter in your own thinking instead of someone else's sentences.

Family and neighborhood complaints

"Parenting by Temperament" (my-child-advisor.lovable.app) is a parenting coach. You jot down the situation, what you did, and how your child reacted, and AI builds a profile of your child's temperament that it keeps revising as records pile up. When tantrums, sleep, or eating problems come up, you can ask how to handle them in a way that fits your child's temperament. There's also a space for parents to work out their parenting philosophy together and record what they agree on.

"Do You Like Mom or Dad Better" (bloomy-parenting-hub.lovable.app) is for when that agreement doesn't go well. When mom and dad clash over parenting, instead of deciding who's right, it weighs both people's worries and intentions fairly to find what works for your family, keeps the child's safety and development at the center, and suggests a small agreement you can try today.

"Is This a Scam?" (sagiscan.lovable.app) was made for a mom who asks "Is this a scam?" every time she gets a text. Upload a text, KakaoTalk message (Korea's main messaging app), or even a YouTube or news screenshot as a photo or text, and AI points out what's suspicious and what to do. No sign-up needed, and it's free up to three times a day. The screens are full of lines written with the parents' generation in mind, like a tip that long-pressing the input box brings up paste, and a reminder to cover passwords and verification codes before uploading.

Sometimes even your daughter is slow to reply. Mom, ask here first now.

"Home" (jibeuro.lovable.app) was built after its maker learned that Korea's public data portal has an OpenAPI for rescued animals. It shows notices for animals in shelters across the country in one place. You can search for a lost pet, share missing reports and sightings, or adopt an abandoned animal. The name comes straight from the hope that many of these animals make it home.

For neighborhood life, someone made "Clean Restroom Near Me" (simple-md-reader.lovable.app), a map that scores restrooms separately for cleanliness and facilities and recommends the clean, well-equipped ones first.

"Rainy Day Shelter Map" (rainy-day-escape.lovable.app) pulls today's weather, and if rain is forecast it recommends indoor spots near you like museums, exhibitions, and cafes, with payment right in the app through Toss Pay.

"Everyone's Last Train" (makcha.lovable.app) is for people who've missed the last train because they stayed a little longer so as not to kill the mood. One person creates a table and shares the link, everyone joins on their own phone and enters their last train time, and it gathers everyone's last train to show in real time how long until the first round breaks up and when each person needs to leave. The remaining time shows as candles melting on the table. For now you can try it with a demo table, and the first screen says our night goes on and the candles tell each of us our time.

So the first person to leave doesn't have to feel sorry.

"Can I Trust You With the Photos?" (pic-coach.lovable.app) is a smartphone photography test that asks: your phone is a pro, but what about the person holding it? When it's unclear which friend or family member should hold the camera on a trip, send them the link. No login, seven photos, and about three minutes later you get their photo sense score. Score 8 or higher and you can trust them with it. It warns up front that the results may be a bit blunt.

Work and career complaints

"Suju-gak" (suju-radar.lovable.app), roughly "Contract Radar," promises to pick out only the freelance projects worth winning. It tackles a problem for freelance developers and studios who scour Wishket (a Korean freelance marketplace), government startup programs, and open KakaoTalk chat rooms every day looking for clients. Every day it collects the Wishket and K-Startup postings that actually went up that day and tells you only the ones that fit your company and tech stack. All you need is your company homepage or a GitHub address. If you don't have one, you can just pick a role like frontend, backend, app, AI and data, planning and PM, or design. For each posting it shows why you, how many competitors there are, and what number you'd be if you applied now.

It also works the other way: when a client posts a request, matching developers get a real-time alert, and if you post a request as a client yourself, you can try both sides on your own. It only curates and takes no fee, so the 20% a platform would normally take gets split between the client and the developer.

"Ctrategy" (insight-navigator-772.lovable.app) is a website diagnosis tool that lets you prep for a 30-minute meeting from a single URL. It's still a demo. It lays out, with evidence, the gap between what the other company's homepage should be doing and what it's actually doing. It goes through three stages: raw material with numbered sources read only from public pages, an internal assessment that organizes problems, opportunities, and meeting questions by evidence and confidence, and a client-facing report with the internal notes stripped out after approval. A person reviews every stage.

"Persuasion Lab" (convince-craft.lovable.app) started from the fact that even after you build a good service or write a good proposal, there are so many people to convince: your team lead, executives, even the security team. Upload a proposal, spec doc, or GitHub address and AI first asks for the numbers and facts you're missing for the pitch. It then lays out what each audience cares about, such as a CEO, CFO, CISO, or team lead, looks up other companies' cases and statistics to check the facts, and writes a persuasion strategy and expected Q&A for each person. You get the results as editable Markdown, 16:9 presentation slides, and a PDF.

"RoleFit" (position-preview.lovable.app) analyzes your fit for a position and your application strategy when you upload a resume and a job posting. It states from the start that it won't invent experience or achievements that aren't on your resume. Resume files are processed only in the browser, and nothing you enter is stored on a server.

"OneCard" (onecard-make-my-card.lovable.app) builds a mobile business card in three minutes at your own address, like onecard/yourname, showing your contact info, career, and portfolio at once, so you can share it with a single link. The sample cards are for a rehab trainer, a freelance designer, and a neighborhood cafe owner, which shows it was built with people outside office jobs in mind.

"Metacognition Test" (cognition-lab-collective.lovable.app) came from the fact that some companies now test metacognition in hiring. You look at 20 words, first predict how many you'll remember, and then take the test. The gap between your prediction and the result tells you whether you're the overconfident type or the modest type. It takes about five minutes without logging in and follows the calibration measures used in metacognition research. You can send your result card to friends or see where you stand among all respondents. Results are stored anonymously and used only for overall distribution statistics.

"How To Work" (howtowork.lovable.app) is a second site from the maker of "Home." It walks through code that AI wrote in the order it runs, explains the logic step by step, and checks the explanation against the code for evidence. You need to get and enter your own OpenAI or Anthropic API key to use it. Uploaded code is stored only in your browser and sent briefly for AI analysis when you ask a question. Nothing stays on a server or in a database.

These days it feels like you can't develop without AI. But it's hard to know how the code AI wrote actually works. It's an ironic situation where I built it but don't know how it works, so this site explains the logic alongside the code to make that easier to understand.

The maker of "CDSS" (mindful-decision-lab.lovable.app) posted the link saying, "I just threw it together and I'm not sure what it is, but something came out." Open it and you find a decision-support prototype for a 2×2 experiment on reducing cognitive bias: a research tool that handles the cognitive biases that creep in when ruling on technical complaints by weighing the balance of evidence. With sample complaints like a re-review of an EV charger conformity assessment or an appeal of a failed elevator inspection, it filters out emotionally charged wording and stacks up facts for each regulation on both sides, using structure to keep the reviewer from locking in a ruling too early or falling into confirmation bias.

"FedLab" (learn-fl-lab.lovable.app) is a learning site for federated learning, where data never leaves the device, from the basics to the research frontier. It splits levels for people new to machine learning, undergrads, and grad students. You follow a weekly curriculum, take one lesson a day, and at the end of each week take a test that checks understanding instead of memorization. Its tagline is clever: just as data never leaves the device, it won't leave your head confused either. The first week alone covers the basic structure of federated learning, the FedAvg algorithm, the problem of data differing from device to device, and why federated learning alone isn't enough to protect privacy.

Apps that connect people and show off taste

"Gyeol" (gyeol-love.lovable.app), meaning "grain" or "texture," started from a complaint posted in the Discord Complaints Center channel before the event: "a dating app for genuine connections." Instead of a dating app where you judge quickly by photos and stats, it's a service where you get to know each other's values and thinking first. Four questions about texting, fighting, weekends, and the pace of a relationship sum up your dating style, and you open with shared tastes like favorite movies or music. It introduces only two people a day, so there's no endless swiping. Send interest on an answer you like, and if you both accept, a private chat begins.

Profile photos are hidden before a match. The blur fades a little as you talk but never becomes fully clear. It doesn't ask for your real name, phone number, or employer, and instead of a compatibility score it tells you where the two of you overlap. Someone left a complaint and someone else took it and turned it into a product in an hour and a half, so this app is exactly what the Complaints Center set out to do.

"ApeTogether" (apetogether.lovable.app) is an English-language app that finds people to climb or run with wherever you go. You can join nearby sessions or host your own, get matched with climbers, and browse community events and a gym map while filling out your own climbing passport.

"ShowMeLook" (showmelook.com) is an AI fashion styling service. Upload one front-facing photo and the stylist Showmi suggests fashion styles that suit you and helps you build your own lookbook. It was the only one of the 30 published on its own domain.

"Chagugyeong" (chagugyeong.lovable.app), roughly "Car Browsing," shows official prices, specs, and clearly sourced photos of premium cars sold in Korea in one place. It currently lists 72 models. Set your budget cap with a slider, filter by body type and fuel, then look at price and spec differences by trim and compare two cars side by side. Every car photo credits the photographer and license terms, and the screen notes that the list is sorted by starting price only and isn't a ranking. You can use it right away with no login or contact info collected. It's still in testing, and the maker said feedback is welcome.

Every price has a check date and a link to the original source, and I didn't include scores, rankings, or recommendations.

Fitting for an event at PUBG Seongsu, someone made "Care Package Draw" (pubgcarepackage.lovable.app): pick a Battlegrounds supply crate, open it, and get a random penalty in an everyday version or a PUBG version.

The maker of "Do You Like Mom or Dad Better" also posted "Reboot My Life" (nextlife-quiz-pix.lovable.app), something to kill time when life feels boring. Answer five odd questions at the Suspicious Next-Life Assessment Office and you get stamped with one of 50 results showing who you'll be in your next life.

A complaints center that doesn't end in a day

The 30 projects went up one by one in the Discord Complaints Center channel between 2:39 and 4:47 p.m., and from 4:00 p.m. we had a demo session to look at them together. From a scam checker for mom to an app that shows your last train with candles to an app that rules on your next life, a lot of them were so inventive it was hard to believe they came out of about an hour and a half.

It was a show-and-tell with no rankings or prizes. The maker of "Gyeol" asked anyone who wanted to try it to hit the heart, and reactions from people who tried things started coming in. The metacognition test got this comment: "I tried it and it's fun. It'd be nice if I could also review how I answered each question." The person who filed a complaint gets something to use right away, and the person who built it gets to keep improving it with users already lined up.

The Complaints Center's next step is to keep improving whatever turns out to be worth using after the event and grow it into AI products people actually use. Every service above is open to try right now, and any feedback you leave goes straight to the makers in the Complaints Center channel on the Bloom Discord (discord.gg/Wd9jaE8MYF). You can post new complaints there anytime, and anyone can file one, even if you didn't come to the event. If reading this made you want to build something yourself, you can start right away at Lovable (lovable.dev).

Thank you to the Bloom members who cleared out a whole weekday afternoon, and to Lovable for giving every participant credits and making this experiment possible. Deep thanks also to PUBG Seongsu, Creative X Seongsu, and the Seongdong Cultural Foundation for the space and the opportunity. The PUBG Seongsu Survivor Hall was a clean space, good for staying heads-down on a laptop all afternoon. The same day, the Tasty Seongsu W.W.C.D chicken-and-beer party was happening outside, so anyone who wanted a break from building could step out for a beer.

Originally written in Korean on Brunch, October 7, 2026.