feat(cursor): show Auto and API pool usage separately#553
feat(cursor): show Auto and API pool usage separately#553AldikYO wants to merge 1 commit intosteipete:mainfrom
Conversation
Cursor Pro plans split quota into two buckets: - Auto pool: base + bonus credits (autoPercentUsed from API) - API pool: base quota only for named models (apiPercentUsed) Previously planPercentUsed (7000/7000 = 100%) was used as primary, making it appear all usage was exhausted even when bonus credits remained. Now primary shows Auto pool and secondary shows API pool. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 7b500b0421
ℹ️ About Codex in GitHub
Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".
| let tertiary: RateWindow? = if let limit = resolvedOnDemandLimit, limit > 0 { | ||
| RateWindow( | ||
| usedPercent: (resolvedOnDemandUsed / limit) * 100, |
There was a problem hiding this comment.
This moves on-demand usage into tertiary, but Cursor’s UI only renders tertiary when supportsOpus is true (checked in MenuDescriptor.accountSection and MenuCardView.makeMetrics), while Cursor metadata keeps supportsOpus false. As a result, on-demand percent/limit is no longer shown as a usage bar (especially when apiPercentUsed is absent, such as legacy/non-token responses), which is a regression from the previous secondary-window behavior.
Useful? React with 👍 / 👎.
| public let openRouterUsage: OpenRouterUsageSnapshot? | ||
| public let cursorRequests: CursorRequestUsage? | ||
| /// Auto + API pool breakdown for token-based Cursor pro plans. | ||
| public let cursorPoolUsage: CursorPoolUsage? |
There was a problem hiding this comment.
The new cursorPoolUsage field is added to UsageSnapshot, but the withIdentity copy path is not updated to pass it through, so any flow that rebuilds a snapshot with a new identity (for example account labeling/scoping) silently drops the Auto/API pool data. That makes the newly added Cursor pool breakdown disappear in multi-account/relabel scenarios even though the fetch parsed it correctly.
Useful? React with 👍 / 👎.
1 participant
Cursor Pro plans split quota into two buckets:
Previously planPercentUsed (7000/7000 = 100%) was used as primary, making it appear all usage was exhausted even when bonus credits remained. Now primary shows Auto pool and secondary shows API pool.