Let’s stop sugarcoating this: Agent v3 is a trainwreck. It’s not a developer tool, it’s a glorified AI-powered loop generator wrapped in buzzwords and sold as innovation. Anyone actually building serious applications can see this thing is architected like duct tape on duct tape.
The issues are not bugs, they’re fundamental design failures:
-
No persistent memory → the Agent can’t even hold context for more than a few steps; try working with a 4000+ file repo and watch it collapse.
-
Execution reliability = zero → audits fail, backend setup is impossible, loops spawn endlessly until your wallet bleeds out.
-
Ghost processes → the Agent literally launches itself and trashes code while burning through credits. That’s not “AI,” that’s a rogue process masquerading as a feature.
-
Billing black hole → one event cost me $51 for a stuck loop, and when $4k+ goes down the drain, your support team blames “user monitoring.” That’s professional gaslighting—it’s your unstable orchestration engine failing, not my oversight.
And yet, all we hear from Replit’s leadership is glossy, performative hype about “AI-native development” and “agents changing coding forever.” No—they’re not. Right now this is an overpriced, underbuilt, half-functioning prototype being marketed as the future.
Let me be clear:
If your Agent can’t persist state, traverse code without choking, or avoid recursive self-destruction, then you have no right calling it a “developer tool.” Developers don’t need trash bots that hallucinate architectures and light money on fire—we need stable systems that earn trust instead of eroding it at 100x cost multipliers.
Replit devs: this isn’t shipping innovation. It’s shipping liability. And every time you push PR spin instead of fixing the orchestration core, you’re proving exactly how disconnected you are from real engineers trying to use this in production.
Cut the interviews, cut the hype. Fix the foundation. Right now, you’re bleeding user trust faster than your Agent burns through compute loops
4 Likes
It is ridiculously expensive. I’m still use replit time to time to test things because I paid the anual membership, but I won’t renew, better and cheaper tools are available.
I tried Agent 3, I asked for a simple prototype for an app, at the end it spend $20 for a buggy prototype, so I stopped it. They are bleeding user trust.
Unfortunately I have the same feedback. I wanted to make a tiny change in my tiny prototype, and it went off on a tangent building a whole robust application/tracking system. Have a look at the exact prompt and interpretation + time and cost. Crazy 
I expected it to be max 2min thinking time $0.18 cost prompt… so it is 50x more indeed…
I hope this gets fixed, and users can have more say before the credits start getting drained on something that wasn’t even asked of in the first place.
Curious what others think and what your experiences are.
Did you chose “Low” on the new autonomy toggle? Try it.
51 minute runtime, you must have been using app testing and/or high or max mode for autonomy. $8.49 for 51 minutes is actually fairly low for that long of a run, in my experience.
@realfunnyeric I don’t have the option you mention. For me it’s just a toggle and it’s been turned off. Print screens attached.
As I said above, you saw the prompt and what it made of it. Something that meant to be a small change agent 3 took as a request of creating achievement tracking system that took 51min and my time to work on that is constrained.
I built the same application with v2 and it definitely wasn’t making up my requests to such extent.
A bit stressed as Monday I’m having time constrained prototyping exercise with Replit and now I’m not sure how to approach it. Would working on a prototype just in design mode help? It seemed to not play this “overachiever” role it does once I worked on updating the built app.
Thanks for help!
Log out, log back in. Let me know.
@realfunnyeric I can see option to adjust. Thank you!
I do see though that it still is going in with full execution of the prompt - so it doesn’t mitigate the issue of it over-interpreting my prompt.
What’s the difference between working with “start with design” and “build the entire app”? If one only wants to build a prototype, should they just stick to iterating within the design mode? Will that make iterating faster?
I’ll be honest, I prototype very little. I build all the working functionality as I go. I haven’t worked in design mode yet, but thanks for the reminder to give it a spin and get familiar.
1 Like
i feel like i have to pay more for solutioning his own errors. WHY???
Horrible. Definitely a huge step backwards for loyal users and a kowtow to the bean counters. Chews thru credits at an insane rate. Supposedly some improvements available and I’ll be picking up a few projects later this week and will see. Certainly felt like it was time to change course after my last few days.
