You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: MEMORY.md
+1Lines changed: 1 addition & 0 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -12,6 +12,7 @@
12
12
- In Pi extensions, do not call action methods such as `getThinkingLevel()` during the factory load; defer them to `session_start` or later handlers.
13
13
- Extension statusline entries should be activity-based: only show an extension in status when it is actively running, retrying, or needs attention; avoid permanent “configured/ready/on” statuses.
14
14
- Codex usage can be queried without Codex CLI by sending Pi's `openai-codex` bearer token to `https://chatgpt.com/backend-api/wham/usage`; response uses Codex `RateLimitStatusPayload` snake_case fields.
15
+
-`pi-codex-usage` statusline must select a rate-limit bucket by current model id/name; `gpt-5.3-codex-spark` can use its own returned bucket instead of primary `codex`.
Copy file name to clipboardExpand all lines: extensions/pi-codex-usage/README.md
+4-1Lines changed: 4 additions & 1 deletion
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -67,9 +67,12 @@ When the selected Pi model provider is `openai-codex`, `pi-codex-usage` refreshe
67
67
68
68
```text
69
69
codex 59% 5h 61% wk
70
+
codex spark 100% 5h 100% wk
70
71
```
71
72
72
-
The statusline value uses the cached usage snapshot and refreshes every five minutes while the current model remains `openai-codex`. Switching away from an OpenAI Codex model clears the item.
73
+
The statusline value uses the cached usage snapshot and refreshes every five minutes while the current model remains `openai-codex`.
74
+
When the selected model has its own returned usage bucket, such as `gpt-5.3-codex-spark`, the statusline switches to that bucket instead of the default `codex` bucket.
75
+
Switching away from an OpenAI Codex model clears the item.
73
76
74
77
Use `/codex-status --no-statusline` for a one-off notification without updating the statusline, or `/codex-status --clear-statusline` to clear the item manually.
0 commit comments