I just wanted to pause for a second and thank the Cursor team for creating such an awesome product. Recently, I’ve noticed some complaints on the forum about pricing, so I felt like sharing why, from my perspective, the $20 subscription is totally worth it.
I previously tested out a few “bring-your-own-key” platforms and quickly realized how expensive it was—I burned through $20 of API use in only five days! With Cursor, though, I’m getting 500 premium requests plus full access to the application—all for that same 20 bucks. To me, that’s an amazing deal.
Another thing that impresses me is how fast the Cursor team rolls out new features and model updates. Their speed of innovation really stands out, and I genuinely think both the product and the team’s hard work don’t get enough recognition.
It’s made a massive difference to my productivity. Now, thanks to Cursor, I can easily put together small micro-apps in just a month, whereas before it would have taken around six months. That $20 monthly investment is actually saving me thousands of dollars in development time and effort.
So, Cursor team—thanks a million! You’ve truly transformed my workflow, and I deeply appreciate your dedication in making something better every day.
Does anyone else feel the same way about the tremendous value Cursor adds to your development process? I’d love to hear your experiences too!
I second your views. For experienced devs, Cursor is a massive productivity boost.
It is the only AI editor for now, which provides slow requests even after 500 fast requests are depleted. A mega corporation like Microsoft is also imposing usage limits at 300 messages in GitHub Copilot for $19 / month without slow requests after that. Compare that to Cursor’s 500 premium requests at $20 / month with slow premium requests.
Then Deepseek v3 is free in Cursor (for both agent + chat).
Cursor lets us add custom docs which helps agents work with new tech and libraries.
The recent global files ignore feature is great for codebase privacy (eg. .env files).
I’m waiting for Supermaven’s integration into Cursor. That should make Cursor tab tab blazing fast.
Cursor has its own low days sometimes where things break. But I’ve noticed the team works fast to fix and innovate. Cursor’s not perfect and it will get better. For now I’m happy with my Cursor subscription.
p.s. I use 3 AI code editors (Cursor, W*sf and the one from Microsoft) to develop 3 projects in parallel. 10x productivity boost. Never been a better time for developers and entrepreneurs.
This is absolutely true, however, I really want the full context window usage (MAX) when I pay 20 dollars a month. I really love MAX, but I don’t want to be paying per request when I’m already paying 20 a month. Hopefully, MAX models will eventually be included in the 20 dollar subscription. If not, I remain a sad developer.
Cursor, if you’re seeing this, please try and unify the pricing so we pay one price for everything and not pay for what we use (it very quickly compounds and stacks up before you know it, especially when the model is also calling multiple tools with MCP servers).
p.s. I use 3 AI code editors (Cursor, W*s f and the one from Microsoft) to develop 3 projects in parallel. 10x productivity boost. Never been a better time for developers and entrepreneurs.
Do you have 1 project per IDE and code in each of them at once?
If you do, do you find the context switching difficult?
Relatively speaking it is cheap indeed. I wonder how they manage to make a profit to be honest, lol. But the bugs lately (“No changes made”) are really killing the product’s reputation. I hope they fix it fast.
With practice I’ve gotten comfortable with the context switching. 3 different projects in 3 editors in parallel can get tricky. But I follow a step by step development workflow for each project. I always start with starter kits (available online or my own). I then add my predefined, project specific Cursor rules. I then create a detailed PRD with ChatGPT o3 or Claude. Only once I have all the project docs in place, do I start developing.
Also if projects are similar in type, then context switching feels easier. (eg webapps or mobile apps). At times these are micro apps which helps too. But if it’s complex software, then all my attention goes to one project. I work using TDD for complex apps so that helps too. As the tests suite grows, I use them as requirements context for agents.
Another big factor is I never vibe code professionally. It maybe good if one has time and money to spend. But i’ve seen it is not sustainable in the long run for complex projects or under financial constraints.
Professional software development was and is still done using established processes, deadlines and various constraints. I treat my workflow as managing a team of junior devs. So the scope, requirements, constraints, deliverables are all laser focused. And we end up creating useable and well tested software within the deadlines.
My workflow is constantly evolving as these model capabilities + editor features grow. But the core elements are grounded in sound software engineering and project management principles. I started professional software development in early 2000’s. Looking back, I feel grateful for the current generation of tools and how much time / effort they save us. And it will only get easier from here on.
Unfortunately the problem to accomplish that is with the Corporate AIs that cursor and many other AI apps use. As long as they only have pay per prompt paying structure, it will unfortunately reflect in Cursor’s pricing structure as well any other AI tools in general out there.
It’s not something Cursor has any control over. It’s why im hopeful that open source coding LLMs will start becoming more accessible and smart. At least the option to add those type of LLMS is already built in cursor. Only then will free prompt limits be possible and paying per prompt will become a thing of the past. But you will either need to own expensive hardware to run the LLMs or subscribe to LLM hosting platforms.