fix(llm): disable json_object mode for :free models on any gateway

The _supports_json_response_format check only looked for
'openrouter.ai' in base_url, missing the Kilo gateway
('api.kilo.ai'). Free-tier models routed through ANY gateway
(Nvidia NeMo :free via OpenRouter OR Kilo) reject json_object
response format, causing the LLM to return content: null.

Error: 'the JSON object must be str, bytes or bytearray, not NoneType'

Fix: check model name for ':free' and 'stepfun/' directly,
regardless of gateway. This covers OpenRouter, Kilo, and any
future gateway.

backend/src/plugins/llm_analysis/service.py

@@ -935,18 +935,15 @@ class LLMClient:
     # @PRE Client initialized with base_url and default_model.
     # @POST Returns False for known-incompatible combinations to avoid avoidable 400 errors.
     def _supports_json_response_format(self) -> bool:
-        base = (self.base_url or "").lower()
         model = (self.default_model or "").lower()
 
-        # OpenRouter routes to many upstream providers; some models reject json_object mode.
-        if "openrouter.ai" in base:
-            incompatible_tokens = (
-                "stepfun/",
-                "step-",
-                ":free",
-            )
-            if any(token in model for token in incompatible_tokens):
-                return False
+        # Free-tier models from ANY gateway often reject json_object mode
+        # (Nvidia NeMo free via OpenRouter or Kilo, stepfun free, etc.)
+        if ":free" in model:
+            return False
+        # stepfun models (even non-free) don't support json_object mode
+        if "stepfun/" in model or model.startswith("step-"):
+            return False
         return True
     # endregion LLMClient._supports_json_response_format