Merge pull request #120 from Fosowl/dev

readme update + avoid asking for clarification multiple times

README.md

@@ -321,7 +321,7 @@ Example config:
 [MAIN]
 is_local = True
 provider_name = ollama
-provider_model = deepseek-r1:1.5b
+provider_model = deepseek-r1:32b
 provider_server_address = 127.0.0.1:11434
 agent_name = Friday
 recover_last_session = False
@@ -339,7 +339,7 @@ stealth_mode = False
 
 - is_local -> Runs the agent locally (True) or on a remote server (False).
 - provider_name -> The provider to use (one of: `ollama`, `server`, `lm-studio`, `deepseek-api`)
-- provider_model -> The model used, e.g., deepseek-r1:1.5b.
+- provider_model -> The model used, e.g., deepseek-r1:32b.
 - provider_server_address -> Server address, e.g., 127.0.0.1:11434 for local. Set to anything for non-local API.
 - agent_name -> Name of the agent, e.g., Friday. Used as a trigger word for TTS.
 - recover_last_session -> Restarts from last session (True) or not (False).
@@ -410,9 +410,12 @@ If this section is incomplete please raise an issue.
 
 **Q: What hardware do I need?**  
 
-7B Model: GPU with 8GB VRAM.
-14B Model: 12GB GPU (e.g., RTX 3060).
-32B Model: 24GB+ VRAM.
+| Model Size  | GPU  | Comment                                               |
+|-----------|--------|-----------------------------------------------------------|
+| 7B        | 8GB Vram | ⚠️ Not recommended. Performance is poor, frequent hallucinations, and planner agents will likely fail. |
+| 14B        | 12 GB VRAM (e.g. RTX 3060) | ✅ Usable for simple tasks. May struggle with web browsing and planning tasks. |
+| 32B        | 24+ GB VRAM (e.g. RTX 4090) | 🚀 Success with most tasks, might still struggle with task planning |
+| 70B+        | 48+ GB Vram (eg. rtx 4090) | 💪 Excellent. Recommended for advanced use cases. |
 
 **Q: Why Deepseek R1 over other models?**  
 

sources/agents/browser_agent.py

@@ -341,7 +341,7 @@ class BrowserAgent(Agent):
                 complete = True
                 break
 
-            if link == None or Action.GO_BACK.value in answer or link in self.search_history:
+            if (link == None and not len(extracted_form)) or Action.GO_BACK.value in answer or link in self.search_history:
                 pretty_print(f"Going back to results. Still {len(unvisited)}", color="status")
                 unvisited = self.select_unvisited(search_result)
                 prompt = self.make_newsearch_prompt(user_prompt, unvisited)

sources/llm_provider.py

@@ -332,6 +332,6 @@ class Provider:
         return thought
 
 if __name__ == "__main__":
-    provider = Provider("server", "deepseek-r1:32b", " 172.81.127.6:8080")
+    provider = Provider("server", "deepseek-r1:32b", " x.x.x.x:8080")
     res = provider.respond(["user", "Hello, how are you?"])
     print("Response:", res)

sources/router.py

@@ -29,6 +29,7 @@ class AgentRouter:
         self.complexity_classifier = self.load_llm_router()
         self.learn_few_shots_tasks()
         self.learn_few_shots_complexity()
+        self.asked_clarify = False
     
     def load_pipelines(self) -> Dict[str, Type[pipeline]]:
         """
@@ -439,9 +440,11 @@ class AgentRouter:
         text = self.lang_analysis.translate(text, lang)
         labels = [agent.role for agent in self.agents]
         complexity = self.estimate_complexity(text)
-        if complexity == None:
+        if complexity == None and self.asked_clarify == False:
+            self.asked_clarify = True
             pretty_print(f"Humm, the task seem complex but you gave very little information. can you clarify?", color="info")
             return None
+        self.asked_clarify = False
         if complexity == "HIGH":
             pretty_print(f"Complex task detected, routing to planner agent.", color="info")
             return self.find_planner_agent()