Uneed - Behind the scenes #1

Hey there 👋🏻

8 years ago, Uneed was just a side project I started during my studies to try Nuxt, a new JavaScript framework. It was called “Uneed frontend” and was a simple, manually curated collection of frontend tools for developers.

Now, it’s been almost 2 years since I started working full-time on it, and it generates between $20k and $30k per month. The stats are getting crazy. More than 300 product submissions are being received daily, 110,000 accounts have been created, 5,000+ people visit the site every day. And it keeps growing!

It’s great, I’m not gonna lie 😅. But it can also be challenging sometimes. And also: it’s not a classic business at all. It’s very different from a SaaS, also very different from a simple directory.

That’s why I’m starting this series of “Behind the scenes”. I want to share some funny, bizarre, surprising short stories about growing Uneed 😊.

Let’s start with the first one:

The slop expansion

It’s a bit risky for the owner of a launch platform to talk about slop software taking over the internet, but here we go!

Anyone can build software now. What used to take months now takes an afternoon, and living from your own products is a dream for a lot of people, so a LOT more people are trying. On Uneed, it shows: around 830 products were added in September 2025, and more than 6,200 in September 2026. October is on track to pass 10,000. That’s 7x more products in a single year, while the number of launches only grew 2.6x.

The gap between those two lines is basically the whole story.

A big part of what comes in never gets to launch day, and what does make it has to wait: right now, more than 7,000 products are scheduled in the waiting line, and the furthest one launches in March 2027 😅.

It’s all about balance: I can’t accept everyone, otherwise it would dilute the visibility AND the backlink value.

Which means I constantly have to find ways to fight this. I won’t disclose everything I did 😅 but one thing is sure: if you just run a directory like this, publish everything without filtering, and give too many backlinks, it won’t work.

What slop actually looks like

“Slop” is a vague word, so here are a few real examples from the database:

• Products built on AI models that don’t exist yet. Veo 4, Kling 5.0, Seedance 4.0… People buy the domain, generate a landing page and submit it before the model is even announced. I counted at least 5 different “Veo 4” websites, and I ended up creating a dedicated refusal reason just for them.
• 23 products in 11 minutes. One Sunday night, someone submitted a metronome, a tuner, a piano, a guitar, a synth and 18 other music tools, each one on a subdomain of the same website.
• The same product, again and again. “AI Image Editor” was submitted 8 times by 8 different people, “Image to Video AI” 7 times and “AI Photo Editor” 6 times. And “test” was submitted 6 times, which is probably the most honest one.
• Calculators and PDF tools everywhere. They’re easy to generate and they can rank on Google, so they multiply: there were almost none in early 2024, and 137 calculators and 106 PDF tools were submitted last quarter alone.

How I sort it out

Reviewing 300+ products a day by hand is impossible, so I built a moderation pipeline. A few simple rules come first (an app store link instead of a website, a page that doesn’t load…), then Claude reads each product page and gives a verdict: keep, refuse, or “I’m not sure”, based on a growing set of rules I’ve defined myself.

When it hesitates, I make the call myself, which happens for about 1 product out of 5, because I told it that refusing a good product is much worse than letting a bad one through. Sooo yes, that’s a lot of work remaining for me 😅.

Why it’s actually good news

If you’re building a real product, all of this works in your favor. Every clone that gets refused is one less product fighting for attention on your launch day, and the people who visit Uneed every day keep coming back because they still find products worth their time.

And if you think we refused your product by mistake, just reply to the email you received. I read every answer, and the AI gets it wrong sometimes too 🙂

That’s it for this first story! In the next one, I’ll tell you about the day I closed the free waiting line… and why I reopened it a few weeks later 🙃

Thomas