feat(chat): add localized error message for model API call and improve error handling in chat history (#947)

src/i18n/locales/en/chat.json

@@ -1,4 +1,7 @@
 {
+    "runError": {
+        "title": "Model API call error"
+    },
     "gatewayNotRunning": "Gateway Not Running",
     "gatewayRequired": "The OpenClaw Gateway needs to be running to use chat. It will start automatically, or you can start it from Settings.",
     "welcome": {

src/i18n/locales/ja/chat.json

@@ -1,4 +1,7 @@
 {
+    "runError": {
+        "title": "モデルAPI呼び出しエラー"
+    },
     "gatewayNotRunning": "ゲートウェイが停止中",
     "gatewayRequired": "チャットを利用するには OpenClaw ゲートウェイが実行されている必要があります。自動的に起動するか、設定から起動できます。",
     "welcome": {

src/i18n/locales/ru/chat.json

@@ -1,4 +1,7 @@
 {
+    "runError": {
+        "title": "Ошибка вызова API модели"
+    },
     "gatewayNotRunning": "Шлюз не запущен",
     "gatewayRequired": "Для использования чата требуется запущенный шлюз OpenClaw. Он запустится автоматически, или вы можете запустить его в Настройках.",
     "welcome": {

src/i18n/locales/zh/chat.json

@@ -1,4 +1,7 @@
 {
+    "runError": {
+        "title": "模型API调用错误"
+    },
     "gatewayNotRunning": "网关未运行",
     "gatewayRequired": "OpenClaw 网关需要运行才能使用聊天。它将自动启动，或者您可以从设置中启动。",
     "welcome": {

src/pages/Chat/index.tsx

@@ -805,7 +805,7 @@ export function Chat() {
 
       {/* Run error callout */}
       {runError && (
-        <div className="px-4 pt-2">
+        <div className="px-4 pt-2" data-testid="chat-run-error">
           <div className="max-w-4xl mx-auto rounded-xl border border-destructive/20 bg-destructive/10 px-4 py-3">
             <p className="text-sm font-medium text-destructive flex items-center gap-2">
               <AlertCircle className="h-4 w-4" />

tests/e2e/chat-task-visualizer.spec.ts

@@ -416,6 +416,9 @@ test.describe('ClawX chat execution graph', () => {
 
       await expect(page.getByTestId('main-layout')).toBeVisible();
       await expect(page.getByText('404 Resource not found')).toBeVisible({ timeout: 30_000 });
+      const runErrorCallout = page.getByTestId('chat-run-error');
+      await expect(runErrorCallout).toBeVisible({ timeout: 30_000 });
+      await expect(runErrorCallout).toContainText('404 Resource not found');
       await expect(page.getByTestId('chat-execution-graph')).toHaveCount(0);
       await expect(page.getByTestId('chat-execution-step-thinking-trailing')).toHaveCount(0);
       await expect(page.getByText('404 Resource not found')).toHaveCount(1);

tests/unit/chat-history-actions.test.ts

@@ -308,6 +308,37 @@ describe('chat history actions', () => {
     ]);
   });
 
+  it('does not set runError from an older assistant failure when a later turn succeeded', async () => {
+    const { createHistoryActions } = await import('@/stores/chat/history-actions');
+    const h = makeHarness({
+      currentSessionKey: 'agent:main:main',
+      runError: 'stale',
+    });
+    const actions = createHistoryActions(h.set as never, h.get as never);
+
+    invokeIpcMock.mockResolvedValueOnce({
+      success: true,
+      result: {
+        messages: [
+          { role: 'user', content: 'first', timestamp: 1773281730 },
+          {
+            role: 'assistant',
+            content: 'fail',
+            timestamp: 1773281731,
+            stopReason: 'error',
+            errorMessage: '400 bad model',
+          },
+          { role: 'user', content: 'retry', timestamp: 1773281732 },
+          { role: 'assistant', content: 'ok now', timestamp: 1773281733 },
+        ],
+      },
+    });
+
+    await actions.loadHistory(true);
+
+    expect(h.read().runError).toBeNull();
+  });
+
   it('retries the first foreground startup history load after a timeout and then succeeds', async () => {
     vi.useFakeTimers();
     const warnSpy = vi.spyOn(console, 'warn').mockImplementation(() => {});