
Anycrap
A digital store that creates fictional products on demand
A year ago Anycrap, a fake store selling things like Dehydrated Water and Anti-Gravity Toilets, hit #1 on HN. Classic viral moment, zero monetization, maximum chaos.

So I did what any sensible person does: ignored it for a year, then over-engineered it.
What I actually shipped:
REST API — 35k absurdist products, free key, filterable by category
CLI — npx anycrap random because terminals deserve nonsense too
faker.js plugin — drop-in for faker.commerce.productName(), works offline, no key needed
HuggingFace dataset — 35k rows if you want to train something questionable
Monetization: a fake product ownership certificate. People actually buy them. I don't fully understand it either.
What's next: no idea. Hoping developers find the faker plugin useful. Or someone trains a model on it. Or nothing happens and I write another update in a year.
I finally did the only responsible thing with 130,000+ Anycrap creations: ran stats on them.
New /stats page: anycrap.shop/stats

A few findings from this extremely serious research:
130,000+ fake products
56,833 unique words used in names
#1 word: “cat”
“cat” beats “dog” 3,141 vs 2,462
1,242 “trump” mentions (zero positive context)
827 “bananas for scale”

8 Likes
Comment
Built anycrap.shop where users can generate any product with AI. They generated everything. Including 1000+ dildos. Manually moved them to a new NSFW category. My life choices are questionable.
What Happened
Gave users unlimited AI product generation.
Expected: fun and quirky absurd stuff
Got: That plus a LOT of adult products
Like, a concerning amount.
The Solution (Manual Labor Edition)
Spent weekends manually reviewing and moving ~1500 products to a new NSFW category.
Yes, manually.
I hope, all of them.
No, I don't want to talk about it.
Current Status
NSFW category: exists
Products: categorized by hand
My browser history: cursed
Automatic filtering: in the backlog
Lessons Learned
1. "Generate anything" is too powerful
The internet will use ALL of it.
2. Manual moderation doesn't scale
This worked for 1000 products. Won't work for 10,000.
3. Should've added filters from Day 1
Hindsight is 20/20. Foresight was 0/20.
4. The internet remains undefeated
Always has been. Always will be.
What's Next
Build actual content filtering before this gets worse.
Keyword detection? AI moderation? Just disabling generation entirely?
Still figuring it out.
Tech: Next.js, Cloudflare Workers AI, regret
Current approach: Manual categorization (send help)
Future approach: Actual filters (hopefully)
Anyone else manually moderate AI-generated content? How's your therapy going?
15 Likes
6 Comments
6 Comments
-
2
This is so freaking funny. Wishing you lots of luck setting up all the filters for workarounds people can find for these things 😂
-
2
This is hilarious and painfully relatable. “Generate anything” really does mean anything. Props to you for pushing through the manual moderation — that’s some real founder grit right there. Definitely learned the hard way that filters aren’t optional!
-
2
Not surprising lol
-
2
Funny how people use powerfull tools for "joy".
-
2
what do you use for image generation ? the image generated are pretty nice.
are the prompts for image generation also generated by AI ?
-
2
It's flux-1-schnell model, and prompt is pretty unique for each product (the random combo of background, camera setup, lighting, product details)
-
The Problem
With 100K+ absurd products in the database, I was drowning. Manually reviewing products meant I can publish maybe 1K/month. At that rate, I'd finish sometime in 2037.
Worse, users kept generating brilliant stuff like "Business Meeting Shortener" alongside complete garbage like "asdasd" or "Bitter Smart" (just... random words with zero intention). I needed a way to separate the genuinely creative absurdity from the actual nonsense.
The Solution: Crap-O-Meter
Built a full AI evaluation pipeline called the Crap-O-Meter that scores every product on three dimensions:
Text Coherence (0-10): Does the description actually develop the title concept?
Image Relevance (0-10): Does the image show what the title promised?
Creativity Factor (0-10): Is this hilariously absurd or just boring?
The system uses multiple Cloudflare AI models in a smart pipeline:
LLM judges for semantic understanding
Embedding models for similarity scoring
Vision models for image analysis
Image captioning with fallback strategies
Technical Highlights (for the nerds)
Bedrock & Pillars architecture: Foundation analysis runs first (gibberish detection, concept extraction), then context-aware evaluations build on that. If the title is just keyboard mashing, kill switch activates immediately - saves ~$0.0015 per garbage product.
Smart caching: Title and description captures are reused across evaluations. No duplicate AI calls. Saves 40-50% on API costs.
Two-pipeline image system: Caption-first approach (fast, cheap) with vision model fallback (slow, expensive). Tries 4 different API payload formats because Cloudflare's APIs are... creative.
Synonym-aware scoring: Knows that "tube" could mean toothpaste container, "bottle" is beer, "alien" is monster. Makes scoring actually work.
The Results
Now every product gets a quality score displayed as an animated gauge (0-100 scale). The system is surprisingly good at:
Catching refusals: "I cannot generate content..." automatic score of 1.0
Rewarding specificity: "Gasoline Powered Turtleneck" scores high, "Faceless Horror" scores low (too vague)
Detecting true creativity: "Cat-To-Human Translator" = 9/10 creativity, "Smart Toaster" = 3/10
Products like "Shit On A Stick For Chefs" now correctly score high (7-8 creativity) because.. it's a good product? Meanwhile, abstract nonsense like random word combinations get brutally low scores.
The Meta-Irony
I'm using AI to judge the quality of AI-generated products on an AI-powered parody store.
It's AIs all the way down.
And honestly It works better than I expected. The system distinguishes between "creative absurdity" and "actual garbage" with surprising accuracy.
What's Next
Using these scores to:
Auto-publish high-quality products (score > 7.5)
Surface the best products for the homepage
Filter the catalog intelligently
Eventually: let users filter by quality scores
Still spending way too much on Cloudflare AI bills, but at least now I can automatically curate the chaos instead of drowning in it.
The most pointless project continues to require increasingly sophisticated engineering. Not sure if I'm optimizing for success or just polishing a very elaborate joke.
2 Likes
Comment
Remember when I had 2000 products and thought that was a lot?
We're now at 115K generated products. This escalated quickly.
Hit #1 on Hacker News, got featured in international tech publications, and watched my Cloudflare AI bill turn into a very real monthly subscription I never asked for. Turns out when you make a viral site where anyone can generate anything for free, they will. Enthusiastically.
The good news: thousands of people enjoyed creating absurd products like "Melatonin Coffee" , "Historical Lego Set: Guillotine", "Barbed Wire Toilet Paper" and much more. The creativity has been genuinely hilarious!
I'm experimenting with ways to keep it sustainable without ruining the fun - turns out monetizing pure absurdity is harder than generating it.
Still the most pointless project, but I like it and keep polishing it with new features. Not sure if that's dedication or delusion.
3 Likes
Comment
Hit 2000 products this week and honestly surprised by how much people are enjoying this ridiculous project. Turns out there's real demand for browsing imaginary garbage products!
Thanks to everyone sharing absurd links and subscribing to our "weekly digest of things you can't buy". The positive feedback for this completely useless site has been amazing.
Still the most pointless project ever, but apparently that's exactly why people love it!
10 Likes
5 Comments
5 Comments
-
2
Haha, I love this. Congratulations on hitting 2000 products you can't buy!
This is marketing genius. You've found a powerful niche in 'anti-productivity.' It's all the fun of discovering weird things online with none of the consumer guilt. The perfect pointless project for our times.
Keep up the amazing work!
-
2
Okay, I have to ask ......out of 2000, which one is your favorite ridiculous product so far?
-
2
I think people like this project because it's quite funny.
Just launched anycrap.shop, a store where you can find and "buy" any product you can possibly imagine.
(You can't actually buy anything, it's all made up. Also, not every product can be found due to AI model limitations).
Type in any name, and our multiverse search engine will find a page for this product with an image, description, and specs.
Currently, the site is automatically being populated with new products, and customers from parallel dimensions are automatically leaving reviews.
I built it in a week mainly to learn technologies - Next.js, Cloudflare Pages, AI content generation. There's not much practical use except sending links to absurd products to friends for lulz or entertaining yourself by typing anything into the search and getting ridiculous and funny results.
Since this isn't actually useful to anyone, I added a subscription feature so everyone who signs up can receive a weekly digest of new arrivals to the store that they'll never be able to purchase. Cool!
Due to model limitations, not all products turn out great, but that even adds to the charm or cringe.
Would be awesome if someone likes this. I'd also be happy to hear any ideas on how to make the site more useful, if that's even possible.
7 Likes
2 Comments
2 Comments
About
I built Anycrap just for fun, thought it would be hilarious to send friends links to "specific crap"










2 Comments
I cannot believe people are buying the fake ownership certificates x)
Why don't you monetize it with ads? it could bring money just by going viral on reddit or even tiktok
I tried Google Ads but they rejected it because of low-quality content. Since then, I'm too lazy and too busy managing other projects