Wherever you code (online, CLI, or IDE), everyone has some love, hate or horror vibe coding stories, from building a prod-ready app in minutes to the whole repo got wiped. Would love to hear your story.

I have been observing numerous posts on social media about just vibe coding your applications and sell it to the client or launch it, People actually ignore the technicality of saas products.

I just want to know from them -

• What about database and its schema, how are you going to store data?

• What about application security?

• What about if you have to just add a new button in your page, are you gonna just ask the AI by pasting the whole code?

• What about scalability of the application?

• Are you sure, your application is not breaking from anywhere?

Been using both for about two days now, and I feel like Fable 5 is actually performing better than Fable 5.1 for me. Orchestration and correctness feel noticeably stronger on Fable 5. Is anyone else seeing this, or is it just my use case? Curious to hear how others are experiencing the two models so far.

When I ask to develop BE and FE everything is going smoothly and better then I can do manually but just I want to develop and agent with another agent - the circus with monkeys comes in my town.

By agents ( developers) I mean Codex , Claude Code ( I tried KIMI Code as well ) and by other agents I mean , well, smth similar but running as web app. So these guys ( agent developers ) totally ignore my instructions and always tend to "test and harness" rather then implement what I am asking for, with maniac determination limit agentic routing with caps, input / output validators and similar stuff.

I built a whole SaaS landing page with auth and pricing in one prompt and felt like a genius. Showed everyone, tweeted about it, the whole thing.

Then I went back to fix one small bug- the signup button. Instead of just fixing it, the AI assumed I meant a completely different button and changed that instead.
When I corrected it, it "fixed" the issue by rewriting a chunk of the auth logic I never asked it to touch- and broke login entirely in the process.
By the time I pointed that out, it had made three more assumptions to "fix" its own fix, and none of them matched what I'd actually built two hours earlier.

I ended up scrolling back through the entire conversation trying to reconstruct decisions I'd already made, just to get back to square one.

Has anyone actually found a real fix for this, or is everyone just quietly rebuilding the same project for the fifth time?

I ve built two marketplaces before:

• One for chatbot templates in 2018 that was used by 700+ marketing agencies.

• Another in 2020 for job seekers in the U.S., which reached 5.5M users.

So let s suppose I have some experience. You can read about me on Bootstrappers, TechCrunch, Dev to.

Now I m considering building a marketplace where creators can list their vibe-coded projects along with the code, a live demo link, etc. The idea is to target:

Ran into this a bunch while vibecoding with longer agent runs.

The agent confidently starts building something a teammate already tried months ago and closed the PR on. It has no way to know, because git only keeps what got merged. Every rejected attempt, and the reason it was rejected, just isn't there. So the agent happily walks into the same wall.

The reasoning behind a codebase (why the timeout is 30, why an approach got dropped) lives in closed PRs and reviews nobody reads again. Your agent can't see any of it.

How are you all handling this? Do you paste past decisions into context by hand, keep a decisions doc, or just catch it in review? Genuinely curious what's working, because review-only has not been enough for me.

I recently switched main agents from Claude Code to Codex, and wow, the code quality feels way higher. But I'm noticing that Codex doesn't explain its decisions or reasoning as clearly as other models. Is it just me, or does Codex skip the 'why' behind the code more often?

Are there tricks to codex out more reasoning? Curious if anyone else has noticed this or found a good workaround.

Hi ProductHunt community,

This term has been coined by someone and there are already more than 80 products that you could put in this category. Looking at the numbers, it's growing pretty fast.

I've been thinking about something.
We've gotten really good at using AI to generate working code, but we're not treating it like production code in terms of documentation.

Traditional developers spend significant time documenting their code because they know future them (or their teammates) will need to understand, modify, or debug it later. But with AI-generated code, we often just copy-paste and move on.

The problem:

You generate a perfect solution today.