diff --git a/apps/api/src/services/openclaw.py b/apps/api/src/services/openclaw.py
index 8348b5d3..448f92f0 100644
--- a/apps/api/src/services/openclaw.py
+++ b/apps/api/src/services/openclaw.py
@@ -343,7 +343,13 @@ class OpenClawService:
             )
             return None
 
-    # AI Provider Implementations - Enhanced with Structured Output
+    # =========================================================================
+    # [ARCHIVED Phase 24 B4 — 2026-04-03 ogt]
+    # 以下三個方法 (_call_ollama/_call_gemini/_call_claude) 為舊版 fallback chain
+    # 新路徑: USE_AI_ROUTER=true → _call_with_fallback → AIRouterExecutor (ai_router.py)
+    # 新 Provider 實作: apps/api/src/services/ai_providers/ (OllamaProvider/GeminiProvider/ClaudeProvider)
+    # 回滾保留: USE_AI_ROUTER=false 時仍由 _call_with_fallback (line ~993) 呼叫此區塊
+    # 完整移除時機: Phase 24 完整驗收後 (ADR-052 D11)
     # =========================================================================
 
     async def _call_ollama(self, prompt: str) -> tuple[str, bool]:
@@ -549,8 +555,12 @@ class OpenClawService:
             logger.warning("claude_call_failed", error=str(e))
             return str(e), False
 
-    # 2026-03-29 ogt: _call_nvidia 已移至 nvidia_provider.py
-    # 符合模組化規範 - 所有 NVIDIA API 呼叫統一由 NvidiaProvider 處理
+    # 2026-03-29 ogt: _call_nvidia 已移至 nvidia_provider.py (ARCHIVED)
+    # 符合模組化規範 - 所有 NVIDIA API 呼叫統一由 NvidiaProvider / OpenClawNemoProvider 處理
+
+    # =========================================================================
+    # [END ARCHIVED Phase 24 B4]
+    # =========================================================================
 
     # =========================================================================
     # Mock LLM - Intelligent Fallback with SignOz Data