{"id":19776,"date":"2026-06-29T15:45:42","date_gmt":"2026-06-29T14:45:42","guid":{"rendered":"https:\/\/innovatenews.site\/index.php\/2026\/06\/29\/i-went-looking-for-the-most-useful-things-people-built-with-claude-code-and-i-found\/"},"modified":"2026-06-29T15:45:47","modified_gmt":"2026-06-29T14:45:47","slug":"i-went-looking-for-the-most-useful-things-people-built-with-claude-code-and-i-found","status":"publish","type":"post","link":"https:\/\/innovatenews.site\/index.php\/2026\/06\/29\/i-went-looking-for-the-most-useful-things-people-built-with-claude-code-and-i-found\/","title":{"rendered":"I went looking for the most useful things people built with Claude Code, and I found\u2026"},"content":{"rendered":"
\n

The topic I went looking for the most useful things people built with Claude Code, and I found\u2026<\/strong> is currently the subject of lively discussion \u2014 readers and analysts are keeping a close eye on developments.<\/p>\n

This is taking place in a dynamic environment: companies\u2019 decisions and competitors\u2019 reactions can quickly change the picture.<\/p>\n

LLMs being able to spit out entire programs that once took teams days or months of hard work to create is frightening. That’s simply something you can’t argue about, and frankly, it’s a little unsettling to watch. However, I don’t think there’s anything good that comes from writing off these tools and refusing to use them. They’re here, and they are only going to get significantly better at what they already do. Something I’ve been talking about for quite a while now is that one of the biggest benefits of vibe-coding is constantly overlooked: the fact that it lets ordinary people build tiny tools for an audience of one.<\/p>\n

Think about it. If you had a crazy app idea, or simply an idea for a tool you wish existed to solve that one specific annoyance that bugs you and you only, you had three realistic options up until a few months ago. You either commit to learning how to code, pay someone who knows how to code to build it for you, or let the idea die in your notes app.<\/p>\n

With AI-assisted coding tools like Claude Code, Replit, Lovable, Codex, and so on, and what they call “vibe-coding” now, you spend an afternoon building it yourself. We, at XDA, have talked a bunch about tools we’ve built for ourselves to scratch our own itches. I decided to flip it around today and see what everyone else was making. So, I went digging through what the Claude community has been building, and compiled all the ones I found impressive. While they’re not all necessarily tools I’d want to replicate myself, they’re worth a look.<\/p>\n

Reddit user Finrandojin built Alexandria, an open-source tool that turns any book or novel into a fully voiced audiobook, and it’s polished enough that it already has nearly 700 stars on GitHub. You feed it a .txt, .md, or .epub, an LLM chunks the text and annotates every line with a speaker and a voice direction, and then a local TTS engine (Qwen3-TTS) performs the whole thing as a multi-voice cast. You can even clone voices, design new ones from a text description, and export to MP3 or a chaptered M4B for your audiobook player.<\/p>\n

The tool runs entirely on your own machine. It works with any OpenAI-compatible LLM, which means you can point it at a local model through LM Studio or Ollama and never send a single page of your book to the cloud, or pay anyone a cent. This is the kind of project that, a couple of years ago, would’ve been a startup with a waitlist and a subscription. Now it’s a free GitHub repo built by one person with Claude, and you can run it tonight!<\/p>\n

\"\"<\/figure>\n

If you’ve ever tried to crochet before, you know the patterns read like a different language. They’re often badly written, crammed with abbreviations, and not even slightly beginner-friendly. Reddit user knowbit hit exactly that wall after picking up crochet to spend less time on his laptop, so he used Claude Code to build a web app to fix it!<\/p>\n

You upload a digital pattern, it rewrites the whole thing in plain words, and then walks you through it step by step. In his comment, he shared that a few of his friends who crochet saw it and started asking for login details. He set out to make one thing for himself, and accidentally built something other people wanted.<\/p>\n

I love to go on walks. Unfortunately, I don’t live in a walkable city, and since there’s no gym near me, I figured the best option was to get a walking pad. The problem with most of them, though, is the software. They come bundled with a proprietary app, a clunky remote, or both, and you’re stuck using whatever the manufacturer decided to ship. Reddit user pdawes had exactly this issue with his generic Chinese desk treadmill, so he used Claude Code to build his own Bluetooth remote app instead.<\/p>\n

This isn’t what I found cool about his project, though. He wired in the Google Maps API so he can “walk” between real destinations and watch his progress on a map or in Street View as he goes. A boring desk treadmill turned into a way to virtually stroll through anywhere on Earth! This is something I’ll be building for myself too, and you bet I’ll be traveling the world from my apartment, one walking session at a time.<\/p>\n

Reddit user horserino built something almost everyone has wished for at some point: a shared shopping list that sorts items by aisle, in the exact order he walks through his store. He shared that they used to manage this with a Google Sheet, which they describe as pretty cumbersome, and decided to vibe-code a replacement using Claude Code. While he didn’t share the app he built (which, unfortunately, a lot of people in these threads don’t), the idea itself is simple enough to picture: a pre-made list, already grouped into the order he walks the store, that he just checks items off into the next shopping run.<\/p>\n

The creator of the app calls it “much dumber” than it sounds and explains that there’s no clever dynamic sorting that hooks into store layouts or anything fancy like that. In fact, the only place AI does any live work is a natural-language input for adding new items, which he himself writes off as a “fancy extra.” That’s it.<\/p>\n

\"\"<\/figure>\n

Now, you could point out that there are already apps on the App Store that already do aisle-sorting (similar to how people in the thread did), and that’s exactly why I picked this somewhat simple app for this list. The point was never that nothing as it existed. Instead, the point is that he didn’t have to settle for the generic version, bend his routine to fit how some app decided a store should be laid out, or pay for features he’d never use. He built the one that matched his store, his order, his household, in an afternoon.<\/p>\n

Not every project here is a quiet little utility. Some of them are gloriously petty, and this is the one that proves it. A user named Pankaj (@the2ndfloorguy on X) hooked his WHOOP band up to his work calendar to figure out, once and for all, which coworker stresses him out the most. Using Claude’s Fable model, he reverse-engineered WHOOP to extract his per-minute heart rate, then matched the spikes to his calendar events and to who was in the room for each one.<\/p>\n

The result is a ranked leaderboard of his colleagues by how much they raise his heart rate! It’s the kind of thing no company would ever build, no manager would ever approve, and frankly probably shouldn’t leave his laptop, which is exactly why it’s brilliant. He took a fitness tracker, a calendar, and a vague workplace grudge, and turned them into actual data.<\/p>\n

Reddit user curious_shawnyv built what he calls a “Tinder for dinner” app for him and his wife: a way to swipe through meal options and land on something they both feel like eating, instead of running the same “what do you want?,” “I don’t know” loop every single night.<\/p>\n

Both he and his wife have ADHD, and as he explains it, they tend to forget their options exist unless they can physically see them. So, the app is essentially an external memory aid. He built custom categories that match how they actually eat (takeout, sit-down, reheating from the freezer, or cooking from scratch) and a scheduler that determines which categories appear by default on any given day. None of that is something a generic meal-planning app would ever bother with, because it was designed around one household’s specific brains and habits.<\/p>\n

If you look closely at these tools, you’ll notice they’re designed to solve niche problems and are built to be customized for the user. And because the person building it is the same person using it, the tool fits their needs exactly. That’s what I wanted this piece to demonstrate.<\/p>\n

Off-the-shelf software is always a compromise. At best, it can get close enough to what you have in mind. But unless the developer can peek into your mind, there will always be a gap. There will always be a feature you wish it had, the option it forces on you, the one annoying thing you’ve learned to live with. When you build it yourself, “close enough” stops being the ceiling. And the barrier to doing that used to be years of learning to code. Now it’s an afternoon and a clear idea of what annoys you.<\/p>\n