As a Pro plan subscriber, I am forced to open this topic due to the serious discrepancy between the platform’s marketing promises and the service actually provided. I am currently unable to use a service that I have paid for.
The Problem: On the Pro plan, for which I pay a regular monthly fee, I am unable to use the Agent, which is a core feature. I am constantly met with the “You’ve hit your usage limit” error. This means I am being arbitrarily deprived of a service for which I have paid.
Deception and Lack of Transparency: Vague and intentionally ambiguous terms like “extended agent limits” are used in marketing materials. However, the reality I’ve encountered is that this is not a “limit” but a hidden “quota” that was never clearly disclosed to me.
The expectation for a “Pro” subscriber is uninterrupted service, not hitting a usage wall mid-month and being forced to pay more. This marketing practice is misleading at best, and an outright deception at worst.
My Demands: Therefore, I demand immediate and straightforward answers from the Cursor team to the following questions:
What is the actual API usage limit for the Pro Plan? Why is this critical information not stated with full transparency and in bold letters on the purchase screen?
Is it your company’s official business model to lure customers into a “Pro” subscription, only to restrict a core function and push for extra payments?
What solution do you plan to offer to remedy the situation for existing users who are unable to access the service they have paid for?
I must emphasize that I expect a satisfactory, honest, and official explanation on this matter.
I support yunnu’s call In my daily work with Cursor, it is clear that Auto Mode does not give me the same satisfactory results as my preferred model Claude-Sonnet-4 (without Max). And for months I was able to use Claude-Sonnet-4 with my Pro plan, but in July, or since one of the last Cursor updates, the error “You Hit the Limits” appears and I am forced to switch to Auto Mode. In contrast to previous months, you can no longer really see in your own dashboard on cursor.com which models are included in the Pro model and for which additional costs are incurred. This is not transparent, doesn’t seem to be fair, and is frustrating in daily work. Looking forward to an honest answer by the Cursor team
Marketing your plans as “unlimited”, yet blocking all requests once an undisclosed quota is reached, is a deceptive practice that forces customers onto higher‑priced tiers. This behaviour amounts to deliberately misleading consumers and profiting through misrepresentation.
If you do not resolve this issue promptly and transparently, ensuring genuinely unlimited access for Pro subscribers, we will pursue legal action, either individually or collectively.
aws just tried the short squeeze on them with kiro and then did a full block.. they’re in crisis mode.. and i’m back to cursor because it still has slightly better interface…
the recent increase in charges and lack of slow requests is a bummer tbh… i assume its all cost based on these newer models now.. idk.
vibe coding is tough. need to get over my 1 shot prompt mind virus. haha
I strongly support the author’s point of view. As paying users, we deserve to receive the services we paid for. If there are limitations or a lack of clear communication, it genuinely feels deceptive. I hope the official team can provide a reasonable explanation and resolve this issue as soon as possible to protect users’ rights.
Yes, I am facing the same issue. The issue is worst for me. I have started subscription on 14th and its been just 3 days. I am getting connection errors, usage limit warnings. Agents are very slow at some point. Not able to generate responses. Initially I thought it is my connection issue but no, I see the pattern, whenever we use cursor for long time, the agent response becomes slow.
Being a paid user, we expect to have service for entire month. They are highly deceptive attracting more users by saying tricky unlimited promises which are not actually
Just use auto. They cam’t tell you limits because the limits change based on supply and demand. They had to change away from unlimited slow requests for all models because as a company they do have to make money eventually. Most customers were getting thousands of dollars of api usage for 20 bucks. If %=ey didn’t make a change they would have shut down after running out of money. Llm’s are very expensive to run. they said it was unlimited but they also had that question mark there that said not all models are unrestricted. Auto is just as good as any other individual model. Garbage in garbage out. sonnet just happens to be a really good ■■■■ polisher.
We are all suffering from this issue, prior plans were transparent and we knew what have used and what we still have in our plan, but in the current plans and awkward terminology, it is intentionally used to confuse the users, which is stressing us to look for an alternative AI coding platform.
In case you didn’t notice, I’m literally mocking the team’s screw-up with the switch to the new pricing model.
But yeah, at least I actually understand how the agent works instead of asking to “turn off cache reading” or whining that “Opus burned through my $20 in less than a month” like some other folks.
And I was already thinking about optimizing my costs back when there was that old system with 500 premium requests.
By the way, I have a feeling that the way Claude interacts with Cursor is kinda buggy, and there might be more requests going back and forth between them than there should be. So it’s worth thinking twice about using it.
In my daily work with Cursor, it is clear that Auto Mode does not give me the same satisfactory results as my preferred model Claude-Sonnet-4 (without Max). And for months I was able to use Claude-Sonnet-4 with my Pro plan, but in July, or since one of the last Cursor updates, the error “You Hit the Limits” appears and I am forced to switch to Auto Mode. In contrast to previous months, you can no longer really see in your own dashboard on 
lol bots by cursor 