I accidentally hit $2.4k MRR using my failed vibe-coded tool, it's kind of funny how it happened

Theory Of Everything

Junior Member
Joined
Apr 18, 2020
Messages
150
Reaction score
158
Hey guys,

Wanted to share something that happened to me recently because it's one of those weird "fail forward" stories that actually worked out.

A while back I threw together this little vibe coded app. The idea was simple: drop your URL, it finds relevant Reddit subs and Facebook groups, then spits out a post you can copy-paste that sounds natural for that community. Nothing too salesy, just subtle. The backend was really technical though, I had to prompt it right, I didn't just have a simple prompt like "find subreddits for URL", I wanted it perfect.

I made the Reddit part free (shows subs + ready posts), and charged $5 one-time to unlock the full thing with Facebook groups. Sounded decent in my head. ...yeah, nobody bought it. Zero sales. I pretty much forgot about it and moved on.

I started working on a different project a few weeks prior as my main idea, it was another vibe-coded app, but this one in a really specific local business niche. Not the usual stuff everyone builds (habit trackers, booking CRMs, etc.). This was genuinely narrow and underserved. I was promoting it only through cold emails and it was moving, but slowly. I wasn't expecting to see any movement for at least 2-3 months.

One evening I was bored and thought, "what the hell, let's see what my dead tool says about this app." I pasted the URL into the dead vibe code I made.

What came back surprised me. It found some really targeted subreddits I had never even heard of, ones with 5k-10k members, very on-topic. Same with a few Facebook groups. The posts it came up with actually sounded like something a real user would write. Brand name mentioned naturally, no hard sell, no AI emoji or that hyphen that it does often.

I figured why not , I copied a couple, posted them, and just waited.

No bans, no deletions. A few people replied or DM'd me asking how they could use the tool for their own situation. I just replied with the free trial link to the new app.

In the first week I got over 35 trial signups. By the end of the month, 6 of them had upgraded to the paid plan at $400/month.

So yeah... $2,400 MRR basically fell into my lap from channels I wasn't even using before lol. All because I fed my failed project into itself.

It's funny, I built that tool thinking it would be the product, but it ended up being way more useful as a helper tool for the next thing.

Made me realize that a lot of these quick vibe-coded experiments aren't really failures if you keep them around. They can become little shortcuts or research weapons for whatever you build next.

I still only had cold emails as my main channel before this. Never really touched Reddit or these niche Facebook groups the right way. Turns out there are some hidden pockets online that are perfect for certain niches if you post the right way (natural, not spammy).

Anyway, just thought I'd share because I've seen a bunch of people here talking about building fast with AI tools lately. Sometimes the thing that flops becomes the best supporting actor for the money project.

Has anyone else had something like this happen? Where a failed tool or experiment ended up helping launch something that actually made money?

Curious to hear your takes.

Moral of the story: You can literally vibe code your own marketing Jarvis who does everything for you from niche research, SEO, market analysis, content writing, image generation, literally everything. Not every vibe coded app needs to be a $10k/m giant, it can be a $0/m gremlin that powers up your main project .
 
Nice share , I m interested in testing the failed tool if it is still available out there !
 
Omg that is honestly such a genius move, like using your own failed project as a secret marketing weapon is totally big brain energy, it just proves that even the random stuff you build can end up being super useful for your main hustle.
 
That’s a great success story. Just because your first app didn’t gain traction doesn’t mean it was a bad app—from what you’ve described, it’s actually a really good, niche app, and the fact that you were able to build it using AI is a major achievement. I wish you the best of luck o
 
it's wild how your 'failed' vibe-coded tool became your secret weapon for finding those hyper-niche communities and totally crushing it with organic reach for your new app
 
Congratulations on the success you deserve it. I have seen a lot of Reddit marketing tools but not so much for Facebook groups, does your tool bring the public ones only or even the private ones, and does it search for groups multi language or EN only? Curious to know and try.
 
Come on, admit it — you found this post through that program and casually posted it here, huh? :)
Shhh dont let it out lol. But no for real I might add that as a feature, forum digging for niche keywords, it'd work great for obscure niches
Omg that is honestly such a genius move, like using your own failed project as a secret marketing weapon is totally big brain energy, it just proves that even the random stuff you build can end up being super useful for your main hustle.
yep exactly
That’s a great success story. Just because your first app didn’t gain traction doesn’t mean it was a bad app—from what you’ve described, it’s actually a really good, niche app, and the fact that you were able to build it using AI is a major achievement. I wish you the best of luck o
Thanks
it's wild how your 'failed' vibe-coded tool became your secret weapon for finding those hyper-niche communities and totally crushing it with organic reach for your new app
yep and i didn't even think to use my own invention until after I needed places to post my new website to get 1000 users
Congratulations on the success you deserve it. I have seen a lot of Reddit marketing tools but not so much for Facebook groups, does your tool bring the public ones only or even the private ones, and does it search for groups multi language or EN only? Curious to know and try.
public and private, some require APIs so I try to use public ones for now
 
Back
Top