Use current-window spend for the OpenRouter Key Limit meter - #101
Merged
Conversation
mstallone
added a commit
that referenced
this pull request
Aug 16, 2026
**TL;DR** — Records the v0.8.11 changelog ahead of tagging the release. **What was happening** - Eight PRs (#101–#108) merged since v0.8.10 — the upstream openusage port batch plus the Gemini 3.7 / Codex long-context pricing fix — with no changelog entry yet. **What this changes** - Prepends the owner-approved v0.8.11 section to `CHANGELOG.md`: seven entries under Bug Fixes, the Sparkle 2.9.5 bump under Chores, and the full commit list for the v0.8.10...v0.8.11 range. **Heads-up** - Once this merges, `v0.8.11` is tagged on the merge commit to start the release run; the same notes are published onto the GitHub Release.
mstallone
added a commit
that referenced
this pull request
Sep 5, 2026
<!-- CURSOR_AGENT_PR_BODY_BEGIN --> ## TL;DR Selective re-implementation of the new upstream OpenUsage commits that still apply on Runway: Codex auto-review stays a first-class model in spend breakdowns and prices as GPT-5.6 Luna from 2026-07-09, and Grok CLI's `grok-proxy` slug resolves to Grok Build. ## What was happening - Upstream has moved on since the last port wave (Runway #101–#108). Each new OpenUsage commit was reviewed against Runway's architecture, existing ports, and the "don't bloat" bar. - Codex still replaced `codex-auto-review` with its dated GPT fallback before aggregation (openusage robinebers#1085), so the breakdown showed gpt-5.5 instead of auto-review. - That fallback table still stopped at gpt-5.5 (openusage robinebers#1125). OpenAI moved auto-review onto GPT-5.6 Luna on 2026-07-09, so every auto-review event since July was priced 25× too high. - Grok CLI now logs `grok-proxy` for Grok Build (openusage robinebers#1123); that slug had no alias, so those tokens were dropped from the spend tiles. ## What this changes - Parser keeps the measured `codex-auto-review` slug and stores the dated GPT fallback on `Event.pricingModel` only for cost. Aggregation keys the rate cache by both slugs so a day that spans the Luna cutoff does not reuse the first event's rates. - Adds `("2026-07-09", "gpt-5.6-luna")` at the top of the auto-review fallback table (ours; ccusage's snapshot still stops at gpt-5.5). Bumps the Codex JSONL cache schema to 2 so already-remapped events are reparsed. - Aliases `grok-proxy` → `grok-build-0.1` and stamps `updated_at` with a same-day ISO timestamp so the feed wins over today's date-only cache. - Documents the auto-review identity/pricing split on the Codex spend-tiles page. ## Heads-up Reviewed and **not** ported, with reasons: - **openusage robinebers#1116 / robinebers#1127** (mandatory daily analytics ping, PostHog bump) — Runway removed analytics in #9. - **openusage robinebers#1111 / robinebers#1106** (scroll / reorder / SVG parse) — Runway already has `ReorderFrameStore`, parsed-once `ProviderMark`, row-local reorder gestures, and the rebuilt popover scroll path. Taking their patch would duplicate and fight that work. - **openusage robinebers#1033** (Reset All Settings) — Runway already has Reset All Customization; a second Settings-wide reset is extra surface, and earlier port waves already skipped it. - **openusage robinebers#1128** (Sparkle 2.9.5 → 2.9.6) — relevant security bump, but this environment cannot refresh `Package.resolved`'s `originHash`. Dependabot is already scheduled weekly; worth a dedicated follow-up. - Pricing/provider commits through robinebers#1112, robinebers#1108, robinebers#1110, robinebers#1109, robinebers#1105, robinebers#1103, robinebers#1101, robinebers#1097, robinebers#1093, robinebers#1089, robinebers#1087, robinebers#1050, robinebers#1019, robinebers#1082 are already in `main`. ## Tests - New parser tests: auto-review lines keep the slug; post-2026-07-09 lines price as Luna; the 2026-07-08/09 boundary stays gpt-5.5. - New aggregation tests: breakdown shows `codex-auto-review` while costing the fallback; Luna and gpt-5.5 auto-review events on either side of the cutoff keep independent rates (guards the rate-cache key). - `grok-proxy` resolves to the same rates as `grok-build-0.1`. - This environment is Linux, so `swift test` could not be run here. CI on `macos-26` is the compile/test gate. <!-- CURSOR_AGENT_PR_BODY_END --> <div><a href="https://cursor.com/agents/bc-28a06551-9894-4f4b-ac34-9d5daf6e62b1?cursor_ref=pr_footer&cursor_cta=open_in_web"><picture><source media="(prefers-color-scheme: dark)" srcset="https://cursor.com/assets/images/open-in-web-dark.png"><source media="(prefers-color-scheme: light)" srcset="https://cursor.com/assets/images/open-in-web-light.png"><img alt="Open in Web" width="114" height="28" src="https://cursor.com/assets/images/open-in-web-dark.png"></picture></a> <a href="https://cursor.com/background-agent?bcId=bc-28a06551-9894-4f4b-ac34-9d5daf6e62b1&cursor_ref=pr_footer&cursor_cta=open_in_cursor"><picture><source media="(prefers-color-scheme: dark)" srcset="https://cursor.com/assets/images/open-in-cursor-dark.png"><source media="(prefers-color-scheme: light)" srcset="https://cursor.com/assets/images/open-in-cursor-light.png"><img alt="Open in Cursor" width="131" height="28" src="https://cursor.com/assets/images/open-in-cursor-dark.png"></picture></a> </div> Co-authored-by: Cursor Agent <cursoragent@cursor.com>
![https://cursor.com/assets/images/open-in-web-dark.png"><source](https://cursor.com/assets/images/open-in-web-dark.png"><source){kind=link}
![https://cursor.com/assets/images/open-in-web-light.png"><img](https://cursor.com/assets/images/open-in-web-light.png"><img){kind=link}
![https://cursor.com/assets/images/open-in-web-dark.png"></picture></a> <a](https://cursor.com/assets/images/open-in-web-dark.png"></picture></a> <a){kind=link}
![https://cursor.com/assets/images/open-in-cursor-dark.png"><source](https://cursor.com/assets/images/open-in-cursor-dark.png"><source){kind=link}
![https://cursor.com/assets/images/open-in-cursor-light.png"><img](https://cursor.com/assets/images/open-in-cursor-light.png"><img){kind=link}
![https://cursor.com/assets/images/open-in-cursor-dark.png"></picture></a> </div](https://cursor.com/assets/images/open-in-cursor-dark.png"></picture></a> </div){kind=link}
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
TL;DR — The Key Limit meter now charges the current limit window's spend against the key's cap, instead of the key's lifetime spend. Port of upstream openusage robinebers#1109.
What was happening
/keyendpoint reportsusage(lifetime spend on the key) andlimit_remaining(what is left in the configured limit window).usageagainst the cap, so once a key had ever spent more than one window's cap, the meter pinned at or past 100% even right after a daily/weekly/monthly reset.What this changes
OpenRouterUsageMapper.keyMetricscomputes used asmax(0, limit − limit_remaining), so the meter tracks the current window.usageof 12 withlimit5 /limit_remaining3 now maps to 2 used of 5.docs/providers/openrouter.mddescribes the window semantics.Heads-up
limitbut nolimit_remainingnow shows as fully used rather than at lifetime spend; this matches upstream's tested behavior, and OpenRouter always returnslimit_remainingalongside a configuredlimit.Tests
swift test --filter OpenRouter— 30 tests, 0 failures.