I was simply trying to add a basic profile picture upload feature to a signup form on my site.
After spending $30 and going in circles for two hours, I switched to high power mode hoping to finally resolve the issue. The agent charged me $14 for just 5 minutes of work—and the problem still isn’t fixed.
Just last week, similar fixes in high power mode cost me $4–$6. That’s a 200% price increase in only 7 days.
At this rate, even multi-billion-dollar corporations wouldn’t pay for it. If I had spent $34 and walked away with a working uploader, fair enough. But I didn’t—the profile picture feature is still completely broken:
Businesses can grow quickly, but they can also decline just as fast.
I genuinely want Replit to succeed, and I’m rooting for them, but the recent changes have me questioning their future.
Was Replit profitable under the Agent 2 pricing model?
If not, does this mean Replit is in a position where profitability is only possible at the current, much higher prices—and if so, what does that say about the long-term sustainability of the business?
I’ve read the posts on the Replit forum with some concern, it’s honestly beyond words. This week I was also surprised by a double layer of invoices for using AGENTV3. For light tasks, nothing too complex, I received two new invoices of $60.60 within just five days. Frankly, I’m quite worried about this development.
I still consider myself lucky that I’m not fully dependent on Replit’s core, since I generate my hosting largely on my own platform (HTML internally, APIs still via Replit). Still, I do plan to take other measures in the coming months.
What really disappoints me is Replit’s attitude. They don’t respond, and that doesn’t align at all with my (or our) way of working. I’ve been working on this project for four months now and I’m only halfway, with still a long way to go. But the cost development is simply unsustainable: from an average of $100 per month to possibly $600 per month or more, plus additional costs once we start exchanging data over Replit’s servers. That’s undesirable, not in line with the times, and for me simply not feasible. At some point, my bags will be unexpectedly packed and left outside (LOL). It’s about time Replit fixed this. Agent 3 doesn’t work much better than 2, that one was much more inventive, faster, and also much more balanced in terms of pricing than is the case now.
I’m curious to see how this will develop. Because if prices keep rising, I’ll have to seek refuge more on other platforms. For example, Lovable, Subabase via GitHub, and I’ll only use Replit for fine-tuning.
I’m currently spending $100 per 8 hours—still cheaper than a good developer, but too much for the results. Agent 3 is slow and introduces bugs, so I’m not sure how you’re spending so little. And forget Replit—it’ll be more expensive than before.
I’ve learned to set up the Agent “strictly” with clear do’s and don’ts. I build my app in WordPress, test everything in WordPress, and make sure the coding aligns with the previous version. Always refer to the correct path/HTML/component. No deploying in Replit, just small steps, and above all, keep it simple.
