Add Codex Chat-to-Responses bridge

- Add Codex provider API format selection and model mapping for Chat-only upstreams.
- Convert Codex Responses requests to Chat Completions and rebuild Chat responses as Responses output.
- Preserve reasoning_content across non-streaming, streaming, tool calls, and previous_response_id follow-ups.
- Add a bounded Codex Chat history cache for restoring tool calls before tool outputs.
- Cover Chat bridge, compact routing, streaming, and history recovery with focused tests.

src/components/providers/forms/CodexFormFields.tsx

@@ -90,6 +90,7 @@ export function CodexFormFields({
 
   const [fetchedModels, setFetchedModels] = useState<FetchedModel[]>([]);
   const [isFetchingModels, setIsFetchingModels] = useState(false);
+  const isChatApiFormat = apiFormat === "openai_chat";
 
   const handleFetchModels = useCallback(() => {
     if (!codexBaseUrl || !codexApiKey) {
@@ -202,7 +203,11 @@ export function CodexFormFields({
               htmlFor="codexModelName"
               className="block text-sm font-medium text-foreground"
             >
-              {t("codexConfig.modelName", { defaultValue: "模型名称" })}
+              {isChatApiFormat
+                ? t("codexConfig.upstreamModelName", {
+                    defaultValue: "上游模型名称",
+                  })
+                : t("codexConfig.modelName", { defaultValue: "模型名称" })}
             </label>
             <Button
               type="button"
@@ -231,13 +236,18 @@ export function CodexFormFields({
             isLoading={isFetchingModels}
           />
           <p className="text-xs text-muted-foreground">
-            {modelName.trim()
-              ? t("codexConfig.modelNameHint", {
-                  defaultValue: "指定使用的模型，将自动更新到 config.toml 中",
+            {isChatApiFormat
+              ? t("codexConfig.upstreamModelNameHint", {
+                  defaultValue:
+                    "Chat 格式下这里填写真实上游模型；接管时 Codex 本地会使用兼容模型，并由路由映射回该模型。",
                 })
-              : t("providerForm.modelHint", {
-                  defaultValue: "💡 留空将使用供应商的默认模型",
-                })}
+              : modelName.trim()
+                ? t("codexConfig.modelNameHint", {
+                    defaultValue: "指定使用的模型，将自动更新到 config.toml 中",
+                  })
+                : t("providerForm.modelHint", {
+                    defaultValue: "💡 留空将使用供应商的默认模型",
+                  })}
           </p>
         </div>
       )}

src/i18n/locales/en.json

@@ -1080,6 +1080,8 @@
     "saveFailed": "Save failed: {{error}}",
     "modelNameHint": "Specify the model to use, will be auto-updated in config.toml",
     "modelName": "Model Name",
+    "upstreamModelName": "Upstream Model Name",
+    "upstreamModelNameHint": "For Chat format, enter the real upstream model here; takeover writes a Codex-compatible local model and routing maps it back to this model.",
     "modelNamePlaceholder": "e.g., gpt-5-codex",
     "contextWindow1M": "1M Context Window",
     "autoCompactLimit": "Auto Compact Limit",

src/i18n/locales/ja.json

@@ -1080,6 +1080,8 @@
     "saveFailed": "保存に失敗しました: {{error}}",
     "modelNameHint": "使用するモデルを指定します。config.toml に自動更新されます",
     "modelName": "モデル名",
+    "upstreamModelName": "上流モデル名",
+    "upstreamModelNameHint": "Chat 形式ではここに実際の上流モデルを入力します。接管時は Codex 互換のローカルモデルを書き込み、ルーティングでこのモデルに戻します。",
     "modelNamePlaceholder": "例: gpt-5-codex",
     "contextWindow1M": "1M コンテキストウィンドウ",
     "autoCompactLimit": "自動圧縮しきい値",

src/i18n/locales/zh.json

@@ -1080,6 +1080,8 @@
     "saveFailed": "保存失败: {{error}}",
     "modelNameHint": "指定使用的模型，将自动更新到 config.toml 中",
     "modelName": "模型名称",
+    "upstreamModelName": "上游模型名称",
+    "upstreamModelNameHint": "Chat 格式下这里填写真实上游模型；接管时 Codex 本地会使用兼容模型，并由路由映射回该模型。",
     "modelNamePlaceholder": "例如: gpt-5-codex",
     "contextWindow1M": "1M 上下文窗口",
     "autoCompactLimit": "压缩阈值",