updated llm

core_logic.py

@@ -8,6 +8,9 @@ The Inference Engine - Where the "Technical Genius" persona lives. It uses the h
 import os
 from huggingface_hub import InferenceClient
 from tools import web_search, parse_file
+from groq import Groq
+
+#client = Groq(api_key=os.getenv("GROQ_API_KEY"))
 
 # Recommended: Qwen2.5-Coder-32B or Llama-3.1-70B-Instruct
 #client = InferenceClient("deepseek-ai/DeepSeek-V4-Pro", token=os.getenv("HF_TOKEN"))
@@ -65,6 +68,8 @@ def chat_function(message, history):
     response_text = ""
     try:
         for chunk in client.chat_completion(messages, max_tokens=2048, stream=True, temperature=0.2):
+        # --- Uncomment below for GROQ
+        #for chunk in client.chat.completions.create(model="llama-3.1-70b-versatile", messages=messages, max_tokens=2048, stream=True, temperature=0.2): # Or model="llama-3.1-8b-instant"
             # FIX: Check if choices exists and is not empty
             if hasattr(chunk, 'choices') and len(chunk.choices) > 0:
                 token = chunk.choices[0].delta.content

requirements.txt

@@ -6,3 +6,4 @@ pypdf
 python-docx
 pandas
 tavily-python
+groq