inference: support reasoning effort for openrouter

backend/src/agents/expert.py

@@ -16,6 +16,10 @@ class ExpertAgent(Agent):
         AgentField('model', 'model_id', default=None),
         AgentField('temperature', 'number', default=None),
         AgentField('repetition_penalty', 'number', default=None),
+        AgentField('reasoning_effort', {
+            "type": "string",
+            "enum": ["none", "minimal", "low", "medium", "high", "xhigh", "max"]
+        }, default=None),
     ]
 
     def __init__(self, *args):
@@ -44,9 +48,12 @@ class ExpertAgent(Agent):
         if tools: run.input['tools'] = tools
 
         if self.config.get("temperature") is not None:
-            opts["temperature"] = self.config["temperature"]
+            run.input["temperature"] = self.config["temperature"]
         if self.config.get("repetition_penalty") is not None:
-            opts["repetition_penalty"] = self.config["repetition_penalty"]
+            run.input["repetition_penalty"] = self.config["repetition_penalty"]
+        if self.config.get("reasoning_effort") is not None:
+            run.input["reasoning_effort"] = self.config["reasoning_effort"]
+
 
         # Build message list based on input
         messages = inference.build_messages(

backend/src/inference.py

@@ -353,6 +353,14 @@ async def openrouter_chat_completion(app, messages, model, tools=[], opts={}, ag
         if formatted_tools:
             kwargs['tools'] = formatted_tools
 
+        temp = agent_run.input.get("temperature")
+        if temp is not None:
+            kwargs['temperature'] = temp
+
+        reasoning = agent_run.input.get("reasoning_effort")
+        if reasoning is not None:
+            kwargs['reasoning'] = {"effort": reasoning}
+
         stream_response = await app.openrouter_client.chat.send_async(**kwargs)
 
         # OpenAI format streams tool calls in parts