Cursor's June 2025 credit switch changed the math for a lot of developers. The Pro plan ($20/mo) now delivers roughly 225 requests with Claude — down from ~500. Run a complex refactor and you're watching credits disappear in real time.
I spent the last few months testing 7 alternatives in actual production work — React, TypeScript, a 25-module ERP system. Not toy projects.
Here's what actually held up...
Top comments (2)
the june 2025 credit change was the moment a LOT of ppl mentally checked out of cursor. honest q: did u try any per-build-priced tools? thats the bet im making w/ moonshift - $3 per shipped SaaS instead of monthly, code generated lands in ur own github + vercel so theres nothing to keep paying for once its shipped. happy to spot u a free first run if u want to add another data point to ur switcher list.
haven't tried moonshift yet, per-build pricing is an interesting model honestly. most of the tools i tested were still subscription based so the math always came down to how many heavy refactors you do per month.
the "code lands in your own github + vercel" part is actually a good selling point, the lock-in with some of these tools is real. might add it as a data point in a follow-up post if i get to test it properly.