What are some impressive vibe coding projects?

DOOMscroll[1] for sure! I still play it since hearing about it on HN.

[1] https://ironicsans.ghost.io/doomscrolling-the-game/

Claude and I created a Stock Ticker app for Windows. https://github.com/krypdoh/TCKR

I vibe-coded myself a flashcard app for latin phrases that I've used quite a lot. https://github.com/eliyastein/latin-learner

"Lessons Learned from Vibe-Coding a Configuration Parser"

https://deanebarker.net/tech/blog/vibe-lessons/

Here's the spec I discuss in the post:

https://deanebarker.net/tech/code/config-lang/spec/

I built a Python script (with a little help from Claude) that scrapes YouTube for football (soccer) match highlights each game week, filters them using regex + Gemini, and emails me the compiled list.

https://ibb.co/chdjvVkQ

I've been building https://smartrik.com which is heavily vibe coded. The entire platform is built out but far from production-ready. Prompts were very compartmentalized and specific so it hasn't been a high level "I want X, build Y" chain of commands, but manually written code has been minimal.

I created a pytorch extension using AI generated code. It wasn’t really created solely off of vibes though as I put significant effort into designing the architecture and all the features, but I had Claude Code write all of the code.

https://github.com/alyxya/mycelya-torch

I created a website and mobile app to display all the coffee shops in the state of Minnesota. It's a .Net Core MVC website and React Native mobile app that is pretty much entirely vibe coded.

I've had fun with building the data loaded, website and mobile app via Claude Code from VS Code. However, I didn't find building the project as enjoyable as actual coding myself. The code is a mess and is definitely overengineered and hard to read. I have had to consistently fix bugs and calibrate my prompts so the machine could produced the features that I was trying to create.

Another thing I learned is to commit early and often. Then create PRs to check what code was updated as things got away from me quickly without me knowing or asking the machine to do the thing it did. A few times entire sections of code were removed that had nothing to do with the feature I was working on. Being able to go back to the previous working commit probably saved me hours.

Vibe coding an entire project was a good experience. There's a lot to learn and focus on the next time I go this route.

Take a look! https://mplscoffee.com

I've created Livi! It is an internal tool for me and my team so we can encode complex images that sometimes have plenty of detail and transparency into optimized AVIF files for our websites.

https://github.com/MARTYR-X-LTD/livi

Backend is Rust and Frontend is SwiftUI. At some point I'll make a Libadwaita frontend for a Linux release. Given my knowledge in Swift and Rust is pretty limited, it was an interesting project to learn the strenghts and limits of LLMs. I've learned quite a lot with it. Most useful lesson is that you might not necessarily need to know the specifics of a language, but you do have to have your common sense skills sound and clear, and how to code architecture a larger project, with refactors here and there, performance optimizations, multithreading, queue, cache and logs, and so on.

Maybe needless to say, but it wasn't easy. The produced code needed to be inspected constantly, and bugfixing, testing, handling edge cases required tons of prompting and guidance. The comparison features, such as pinch to zoom, keep the zoom and image positions while switching between the different generated images, handling exporting, all these features were loaded with intricacies. So far, glad that I was able to produce this.

As a fun fact, now that AVIFENC supports tune=iq, we don't have to mess with specific encoder settings and find the proper quality number anymore, but still, learned a lot from it

WordsUnite.us a fully vibe coded chant synchronization platform. So far, huge failure. But it’s impressive that I didn’t have to write any of the code! Lots of spec writing and validating with Claude Code. Claude wrote it all.

Interesting that this thread mostly consists of people sharing vibe-coded projects they themselves made, not people sharing other people’s projects that they’ve found useful. The latter would provide higher confidence that the project is actually impressive. It’s the IKEA effect for software.

Claude and I (mostly Claude) built a web app for lawyers that uses Gemini to extract (and calculate, where necessary) court dates and deadlines from a PDF scheduling order and output them in a structured format. The user can then select which ones they want to add to their calendar and generate an ICS file to add all case dates/deadlines at once.

https://www.courtsynccalendar.com/

Words unite.us fully vibe coded, only 1 user.

I'm giving network admin courses. All my slides have done with Claude and Chatgpt. Formations.mehdiguiraud.com/cours/

I vibecoded a platform that reads a csv/excel and outputs a few different analysis / charts out of it automatically.

Works quite well

Nothing I can show you but I still think a notable story:

A non-developer (but PM) friend of mine vibe-coded his own strength-training programming app that he uses daily. He pivoted from that to vibe-coding his own 2d game engine and 4x strategy game he's been working on.

He's effectively used vibe-coding to teach himself how to code and he brings me questions/problems as good as any working junior developer.

I created DecayBlock, mainly to help people stay focussed online but also to teach myself JavaScript.

I've been using it for the past few months and it's massively boosted my productivity.

Chrome: https://chromewebstore.google.com/detail/decayblock/lpljcnal... Firefox: https://addons.mozilla.org/en-GB/firefox/addon/decayblock/

Here is my personal project that tracks my Aldi grocery prices of products over time. My rule for the project is, all code must be done by the Copilot agent within VSCode. I did hand type the Git commits.

https://aldi-prices.lawruk.com/ https://github.com/jimlawruk/aldi-prices

I've done quite a few projects with ChatGPT. I just can't vibe (no pun intended) with any of the agent tools or AI plugins etc. Also not having to worry about API costs is the main reason so I can be as spammy as I like.

Site for a drum and bass community. Isn't quite there but nearly - test.dnbfamily.com

Completely vibe coded local LLM note formatter (I broke the latest release, need to fix) - https://formait.app/

Two android apps:

When you can't decide what to watch (swipe left + right like tinder) - https://play.google.com/store/apps/details?id=com.blazingban...

Bluetooth mesh app - https://play.google.com/store/apps/details?id=com.blazingban...

C# app that monitors trading performance of CTrader, locks down the application if your PNL goes too high or low (configurable) so that you don't lose all your money when you're tilted. Built for a friend that does a fair amount of trading.

So many little scripts and stuff for personal or work projects.

----

What are you trying to learn from the videos?

I don't really have anything substantial (outside of some one-shot CLI tools, etc.) since I subscribe to the strict definition of vibe coding as a form of "coding behaviorism" (per Karpathy's original tweet).

However, Piece Together [1] is a large game that I made a few months ago with a decent amount of agentic tooling (RooCode and Claude Code). It lets players solve animated forms of jigsaw puzzles.

[1] https://animated-puzzles.specr.net

I submitted and reply to "Ask HN: What are you working on? (October 2025)" thread 7 days ago and receive out-of-expectation response. So, I will put my project here. Not totally vibe code, I prefer the term AI-assisted.

I'm currently building my own coding agent, VT Code. VT Code is a Rust-based terminal coding agent with semantic code intelligence via Tree-sitter (parsers for Rust, Python, JavaScript/TypeScript, Go, Java) and ast-grep (structural pattern matching and refactoring).

It supports multiple LLM providers: OpenAI, Anthropic, xAI, DeepSeek, Gemini, OpenRouter, Z.AI, Moonshot AI, all with automatic failover, prompt caching, and token-efficient context management. Configuration occurs entirely through vtcode.toml, sourcing constants from vtcode-core/src/config/constants.rs and model IDs from docs/models.json to ensure reproducibility and avoid hardcoding. [0], [1], [2]

Recently I've added Agent Client Protocol (ACP) integration. VT Code is now a fully compatible ACP agent, works with any ACP-clients: Zed (first-class support), Neovim, marimo notebook. [3]

[0] https://github.com/vinhnx/vtcode

[1] https://crates.io/crates/vtcode

[2] https://docs.rs/vtcode

[3] https://agentclientprotocol.com/overview/agents

Thank you!

Not so much project as task, but I have been really happy with "vibe coding" to fix my production crashes. In some cases these crashes are impossible for me to reproduce as a solo, and I trusted the fixes provided by codex which seemed to have worked based on the absence of these crashes now.

Maybe more suitable for a vibe debugging addendum.

https://run-phx.com/ is a guide to trail running in PHX... wrote all the content / reviews, but was happy to let claude handle the NextJS / Firebase backend work. Remarkable what a good job it did, although it was definitely a conversation

https://github.com/HowProgrammingWorks/UUID/ My programming teacher has created his own LLM-assisted UUID implementation to achieve a balance between the time to generate and the bit depth. If I understand correctly this ID implementation consumes the least time per 1 bit of entropy among all JS implementations. It is known that the technical task was three times longer than the code and the code has been generated with just one attempt. The teacher has not modified the output.

I built a pretty janky self-hostable version of https://archive.org called Arker. It's open source here: https://github.com/hackclub/arker

It's hosted at https://archive.hackclub.com. You can see a sample archived page here: https://archive.hackclub.com/lX1mm.

It's archived about 22k links so far and hasn't crashed yet!