diff --git a/app/routers/chat.py b/app/routers/chat.py
index 721a05d..8cb5bc9 100644
--- a/app/routers/chat.py
+++ b/app/routers/chat.py
@@ -1,12 +1,9 @@
 """
 FILE: app/routers/chat.py
 DESCRIPTION: Haupt-Chat-Interface (RAG & Interview). Enthält Intent-Router (Keywords/LLM) und Prompt-Construction.
-VERSION: 2.7.2 (Deep Fallback Edition)
+VERSION: 2.7.3 (Debug & Deep Fallback Edition)
 STATUS: Active
-FIX: Respektiert preferred_provider aus decision_engine.yaml und implementiert 
-     Deep Fallback Logik zur Vermeidung leerer Cloud-Antworten (Silent Refusal).
-DEPENDENCIES: app.config, app.models.dto, app.services.llm_service, app.core.retriever, app.services.feedback_service
-EXTERNAL_CONFIG: config/decision_engine.yaml, config/types.yaml
+FIX: Integriert erweiterte Debug-Logs zur Analyse von Context-Overflow und Silent Refusals.
 """
 
 from fastapi import APIRouter, HTTPException, Depends
@@ -313,9 +310,17 @@ async def chat_endpoint(
             final_prompt = template.replace("{context_str}", context_str).replace("{query}", request.message)
             sources_hits = hits
         
+        # --- DEBUG SPOT 1: PROMPT CONSTRUCTION ---
+        logger.info(f"[{query_id}] PROMPT CONSTRUCTION COMPLETE. Length: {len(final_prompt)} chars.")
+        if not final_prompt.strip():
+            logger.error(f"[{query_id}] CRITICAL: Final prompt is empty before sending to LLM!")
+
         # --- GENERATION MIT DEEP FALLBACK ---
         system_prompt = llm.get_prompt("system_prompt")
         
+        # --- DEBUG SPOT 2: PRIMARY CALL ---
+        logger.info(f"[{query_id}] PRIMARY CALL: Sending request to provider '{preferred_provider}'...")
+        
         # 1. Versuch mit konfiguriertem Provider (z.B. Ollama für EMPATHY)
         answer_text = await llm.generate_raw_response(
             prompt=final_prompt, 
@@ -326,7 +331,9 @@ async def chat_endpoint(
 
         # DEEP FALLBACK: Wenn die Antwort leer ist (Silent Refusal in der Cloud)
         if not answer_text.strip() and preferred_provider != "ollama":
-            logger.warning(f"🛑 [{query_id}] Leere Antwort von '{preferred_provider}'. Starte LOKALEN FALLBACK via Ollama...")
+            # --- DEBUG SPOT 3: FALLBACK TRIGGER ---
+            logger.warning(f"🛑 [{query_id}] PRIMARY '{preferred_provider}' returned EMPTY. Triggering Deep Fallback to Ollama...")
+            
             answer_text = await llm.generate_raw_response(
                 prompt=final_prompt, 
                 system=system_prompt,