Pro subscribers should retain basic features (Tab) + clearer usage metrics (Usage $ vs Spending %)

Hi Cursor team,

I’m a Pro subscriber and want to share feedback on usage limits and how they’re displayed. The current experience has significantly affected how I use the product.

1. Two usage dashboards are confusing

The dashboard shows two different metrics that are easy to misread:

  • Usage (dollar amount, e.g. ~$19.9 “Included”)
  • Spending → Included in Pro (percentage, e.g. 11% Auto, 0% API)

I initially assumed that when Usage approached $20, I was about to run out of my Pro quota. In reality, my Spending was only around 11–12%.

Because of this, I switched to DeepSeek last month even though I still had plenty of included quota left. I only learned later (from forum posts and support replies) that Spending % is the metric that actually controls limits, while Usage $ is informational and can exceed $20 without meaning I’m close to the cap.

Request: Please either merge these metrics, or add a clear, prominent notice on the Usage page, e.g.:

“The dollar amount is an estimate only. Your actual quota is determined by Spending %. Limits apply at 100%, not when Usage $ reaches $20.”

This would prevent users from unnecessarily limiting themselves or leaving Cursor for other tools.


2. Basic features like Tab should remain available for Pro users

I understand that Pro includes a monthly quota rather than unlimited usage. However, I believe basic features such as Tab completion should remain available for active Pro subscribers, even when the included quota is exhausted.

My reasoning:

  1. Tab feels like a core IDE feature, not a premium “frontier model” feature. As a paying Pro user, I expect baseline functionality to remain usable.
  2. Tab requests are relatively lightweight compared to Agent or heavy chat usage, but they are still blocked when quotas hit 100%.
  3. BYOK does not help here. I can use my own API key for manually selected chat models, but Tab, Agent, Auto, and Apply still run on Cursor’s infrastructure and count toward subscription limits. So even when I’m paying my model provider directly, I can still lose Tab—which doesn’t feel fair.
  4. If the concern is cost, a modest baseline allowance (e.g. a minimum Tab quota per billing cycle that is never throttled) would be a reasonable compromise for Pro subscribers.

Request: Consider one of the following for Pro (and ideally Pro+ / Ultra) subscribers:

  • Option A: Tab (and possibly other “basic” features) are never rate-limited while the subscription is active.
  • Option B: A guaranteed minimum Tab allowance per billing cycle, separate from the main Spending pools.
  • Option C: BYOK users get a separate, non-throttled baseline for Tab, since chat usage with their own key does not draw from included API quota.

Hey, thanks for the detailed feedback, especially on the metrics. You nailed it.

  1. Usage $ vs Spending %

You got it right, these are two different metrics.

  • Usage $ is informational. It’s the retail value of the included usage you’ve already consumed. It can go up to $20 or even higher, and that alone does not mean you’re close to the limit.
  • Spending % is what actually controls your quota. The limit hits at 100%, not when the dollars reach $20.

Based on your screenshots, it all checks out. Total spend $19.33 with 9% usage means you’ve only used a small part of your quota. So switching to DeepSeek last month wasn’t necessary, you still had plenty of room. It sucks that the UI made this unclear.

I agree it’s not intuitive. I’ll pass this to the team to make it clearer in the dashboard, or to merge the metrics. This feedback comes up a lot, so you’re not the only one.

  1. Tab for Pro

Good news here. Tab on Pro is already unlimited and does not come out of the spend pool. That means Tab completions keep working even at 100% quota. Tab only gets blocked for Free users. So the baseline you’re asking for is basically already there by design.

If you’re actually seeing Tab stop working when you hit your quota, that’s a bug, not expected behavior. If that happens, send the Request ID right top corner of the chat > Copy Request ID and your Cursor version, and we’ll take a look.

Let me know if any of this doesn’t match what you’re seeing.