v0.9.9

What's Changed

• fix(codex): drop the client-only "ultra" reasoning effort from discovered catalogs by @YanZiBin in #130
• style: 订阅设置和用量浮层对齐原生 UI by @kermars39-web in #124
• perf: serve cached model catalogs and refresh hourly by @kermars39-web in #127
• feat: 支持订阅和模型自定义显示名 by @kermars39-web in #123
• feat(claude): make prompt-cache TTL configurable (5m or 1h) by @prgrmrwy in #131
• feat(codex): display and manually redeem usage-limit resets by @victorwads in #134

New Contributors

• @YanZiBin made their first contribution in #130
• @kermars39-web made their first contribution in #124

Full Changelog: v0.9.8...v0.9.9