Add application files

app.py

@@ -52,16 +52,19 @@ print("🚀 CX AI AGENT - ENTERPRISE B2B SALES INTELLIGENCE")
 print("="*80)
 
 # AI Mode - HuggingFace Inference API (FREE)
-# Uses huggingface_hub InferenceClient
-HF_MODEL = os.getenv("HF_MODEL", "mistralai/Mistral-7B-Instruct-v0.3")
+# Uses huggingface_hub InferenceClient with Qwen/Qwen3-4B
+HF_MODEL = os.getenv("HF_MODEL", "Qwen/Qwen3-4B")
 HF_TOKEN = os.getenv("HF_TOKEN") or os.getenv("HF_API_TOKEN")
 
+# Session token storage (can be set via UI)
+session_hf_token = {"token": HF_TOKEN}
+
 print(f"🤖 AI Mode: HuggingFace Inference API (FREE)")
 print(f"   Model: {HF_MODEL}")
 if HF_TOKEN:
-    print(f"✅ HF_TOKEN loaded")
+    print(f"✅ HF_TOKEN loaded from environment")
 else:
-    print("⚠️  HF_TOKEN not set! Get one at: https://huggingface.co/settings/tokens")
+    print("ℹ️  HF_TOKEN not set in environment - can be entered in the Setup tab")
 
 serper_key = os.getenv('SERPER_API_KEY')
 if serper_key:
@@ -89,7 +92,8 @@ def warmup_hf_model():
     Send a dummy prompt to warm up the HuggingFace Inference API.
     This ensures the model is loaded and ready for the first real request.
     """
-    if not HF_TOKEN:
+    token = session_hf_token.get("token")
+    if not token:
         print("⏭️  Skipping model warm-up (HF_TOKEN not set)")
         return
 
@@ -97,7 +101,7 @@ def warmup_hf_model():
         from huggingface_hub import InferenceClient
         print(f"🔥 Warming up HuggingFace model ({HF_MODEL})...")
 
-        client = InferenceClient(token=HF_TOKEN)
+        client = InferenceClient(token=token)
         # Send a simple dummy prompt to warm up
         response = client.chat_completion(
             model=HF_MODEL,
@@ -116,6 +120,16 @@ def warmup_hf_model():
         print(f"⚠️  Model warm-up skipped: {e}")
 
 
+# Helper function to get current HF token (from UI or environment)
+def get_hf_token(ui_token: str = None) -> str:
+    """Get HF token from UI input, session storage, or environment"""
+    if ui_token and ui_token.strip():
+        # Update session storage with UI token
+        session_hf_token["token"] = ui_token.strip()
+        return ui_token.strip()
+    return session_hf_token.get("token") or ""
+
+
 # Run warm-up in background to not block startup
 import threading
 warmup_thread = threading.Thread(target=warmup_hf_model, daemon=True)
@@ -569,13 +583,19 @@ def reset_all_data():
 # ============================================================================
 # CLIENT SETUP - Research the user's company
 # ============================================================================
-async def setup_client_company(company_name: str, progress=gr.Progress()):
+async def setup_client_company(company_name: str, hf_token_input: str, progress=gr.Progress()):
     global knowledge_base
 
     if not company_name or not company_name.strip():
         yield "⚠️ Please enter your company name."
         return
 
+    # Get HF token from UI input or environment
+    token = get_hf_token(hf_token_input)
+    if not token:
+        yield "⚠️ **HF_TOKEN Required**: Please enter your HuggingFace token in the Setup tab.\n\nGet a free token at: https://huggingface.co/settings/tokens"
+        return
+
     company_name = company_name.strip()
 
     output = f"## 🏢 Setting Up: {company_name}\n\nBuilding knowledge base...\n\n---\n\n### ⏳ Progress\n\n"
@@ -586,7 +606,8 @@ async def setup_client_company(company_name: str, progress=gr.Progress()):
         # Initialize HuggingFace agent (FREE Inference API)
         agent = AutonomousMCPAgentHF(
             mcp_registry=mcp_registry,
-            model=os.getenv('HF_MODEL', 'mistralai/Mistral-7B-Instruct-v0.3')
+            hf_token=token,
+            model=HF_MODEL
         )
         output += f"✅ AI Agent initialized (HuggingFace - {agent.model})\n\n"
         yield output
@@ -754,13 +775,19 @@ This is OUR company - we need this information to find matching prospects."""
 # ============================================================================
 # AI PROSPECT DISCOVERY - Automatically find prospects
 # ============================================================================
-async def discover_prospects(num_prospects: int = 5, progress=gr.Progress()):
+async def discover_prospects(num_prospects: int, hf_token_input: str, progress=gr.Progress()):
     global knowledge_base
 
     if not knowledge_base["client"]["name"]:
         yield "⚠️ **Setup Required**: Please go to Setup tab and enter your company name first."
         return
 
+    # Get HF token from UI input or session
+    token = get_hf_token(hf_token_input)
+    if not token:
+        yield "⚠️ **HF_TOKEN Required**: Please enter your HuggingFace token in the Setup tab.\n\nGet a free token at: https://huggingface.co/settings/tokens"
+        return
+
     client_name = knowledge_base["client"]["name"]
     client_info = knowledge_base["client"].get("raw_research", "")
 
@@ -772,7 +799,8 @@ async def discover_prospects(num_prospects: int = 5, progress=gr.Progress()):
         # Initialize HuggingFace agent (FREE Inference API)
         agent = AutonomousMCPAgentHF(
             mcp_registry=mcp_registry,
-            model=os.getenv('HF_MODEL', 'mistralai/Mistral-7B-Instruct-v0.3')
+            hf_token=token,
+            model=HF_MODEL
         )
         output += f"✅ AI Agent initialized (HuggingFace - {agent.model})\n\n"
         yield output
@@ -1193,6 +1221,21 @@ def create_app():
                     with gr.Column(scale=1):
                         gr.HTML("""
                         <div class="action-card">
+                            <h3>🔑 HuggingFace Token</h3>
+                            <p>Enter your HuggingFace token to use the AI. Get a free token at <a href="https://huggingface.co/settings/tokens" target="_blank">huggingface.co/settings/tokens</a></p>
+                        </div>
+                        """)
+
+                        hf_token_input = gr.Textbox(
+                            label="HuggingFace Token",
+                            placeholder="hf_xxxxxxxxxxxxxxxxxxxxxxxxx",
+                            type="password",
+                            lines=1,
+                            value=HF_TOKEN or ""
+                        )
+
+                        gr.HTML("""
+                        <div class="action-card" style="margin-top: 16px;">
                             <h3>🏢 Your Company</h3>
                             <p>Enter your company name. AI will research your company and then automatically find matching prospects.</p>
                         </div>
@@ -1204,9 +1247,9 @@ def create_app():
                             setup_btn = gr.Button("🚀 Setup My Company", variant="primary", size="lg")
                             reset_btn = gr.Button("🗑️ Reset All", variant="stop", size="sm")
 
-                        gr.HTML("""
+                        gr.HTML(f"""
                         <div class="action-card" style="margin-top: 16px;">
-                            <h3>🤖 Automated Workflow</h3>
+                            <h3>🤖 AI Model: {HF_MODEL}</h3>
                             <ol style="margin: 0; padding-left: 20px;">
                                 <li><strong>Setup</strong> — AI researches your company</li>
                                 <li><strong>Find Prospects</strong> — AI discovers matching companies</li>
@@ -1217,7 +1260,7 @@ def create_app():
                         """)
 
                     with gr.Column(scale=2):
-                        setup_output = gr.Markdown("*Enter your company name to begin.*")
+                        setup_output = gr.Markdown("*Enter your HuggingFace token and company name to begin.*")
 
             # ===== DASHBOARD TAB =====
             with gr.Tab("📊 Dashboard", id=1):
@@ -1231,7 +1274,7 @@ def create_app():
                     prospects_stat = gr.HTML(get_stat_html("0", "Prospects Found", "var(--primary-blue)"))
                     contacts_stat = gr.HTML(get_stat_html("0", "Decision Makers", "var(--success-green)"))
                     emails_stat = gr.HTML(get_stat_html("0", "Emails Drafted", "var(--warning-orange)"))
-                    gr.HTML(get_stat_html("Granite", "AI Model", "var(--purple)"))
+                    gr.HTML(get_stat_html("Qwen3-4B", "AI Model", "var(--purple)"))
 
             # ===== PROSPECTS TAB =====
             with gr.Tab("🎯 Prospects", id=2):
@@ -1248,6 +1291,15 @@ def create_app():
                         </div>
                         """)
 
+                        # HF Token input for prospects tab (linked to setup tab token)
+                        hf_token_prospects = gr.Textbox(
+                            label="HuggingFace Token (from Setup tab)",
+                            placeholder="hf_xxxxxxxxxxxxxxxxxxxxxxxxx",
+                            type="password",
+                            lines=1,
+                            value=HF_TOKEN or ""
+                        )
+
                         num_prospects = gr.Slider(minimum=1, maximum=10, value=3, step=1, label="Number of prospects to find")
                         discover_btn = gr.Button("🔍 Find Prospects & Contacts", variant="primary", size="lg")
 
@@ -1316,7 +1368,7 @@ def create_app():
         # Setup button - run setup and then update status indicators
         setup_btn.click(
             fn=setup_client_company,
-            inputs=[client_name_input],
+            inputs=[client_name_input, hf_token_input],
             outputs=[setup_output]
         ).then(
             fn=lambda: (get_client_status_html(), get_client_status_html()),
@@ -1338,7 +1390,7 @@ def create_app():
         # Discover prospects and then update all lists
         discover_btn.click(
             fn=discover_prospects,
-            inputs=[num_prospects],
+            inputs=[num_prospects, hf_token_prospects],
             outputs=[discovery_output]
         ).then(
             fn=lambda: (get_prospects_html(), get_contacts_html(), get_emails_html()),

mcp/agents/autonomous_agent_hf.py

@@ -81,9 +81,19 @@ class AutonomousMCPAgentHF:
         """
         self.mcp_registry = mcp_registry
         self.hf_token = hf_token or os.getenv("HF_TOKEN") or os.getenv("HF_API_TOKEN")
-        self.provider = provider or os.getenv("HF_PROVIDER") or DEFAULT_PROVIDER
         self.model = model or os.getenv("HF_MODEL") or DEFAULT_MODEL
 
+        # Auto-detect provider based on model
+        if provider:
+            self.provider = provider
+        elif os.getenv("HF_PROVIDER"):
+            self.provider = os.getenv("HF_PROVIDER")
+        elif self.model in QWEN3_MODELS or self.model.startswith("Qwen/Qwen3"):
+            # Qwen3 models need a provider (use nebius by default)
+            self.provider = "nebius"
+        else:
+            self.provider = DEFAULT_PROVIDER
+
         if not self.hf_token:
             raise ValueError(
                 "HF_TOKEN is required!\n"
@@ -94,10 +104,14 @@ class AutonomousMCPAgentHF:
         # Initialize HuggingFace InferenceClient
         try:
             from huggingface_hub import InferenceClient
-            self.client = InferenceClient(
-                provider=self.provider,
-                token=self.hf_token
-            )
+            # For serverless API (hf-inference), don't pass provider
+            if self.provider == "hf-inference":
+                self.client = InferenceClient(token=self.hf_token)
+            else:
+                self.client = InferenceClient(
+                    provider=self.provider,
+                    token=self.hf_token
+                )
             logger.info(f"HuggingFace InferenceClient initialized")
             logger.info(f"  Provider: {self.provider}")
             logger.info(f"  Model: {self.model}")

requirements.txt

@@ -1,9 +1,10 @@
 # CX AI Agent - Requirements for HuggingFace Spaces
 # ================================================
 
-# HuggingFace Hub (for FREE Inference API)
+# HuggingFace Hub (for Inference API with Qwen/Qwen3-4B model)
 # Provides InferenceClient for chat completions
-huggingface_hub>=0.22.0
+# Requires HF_TOKEN for authentication
+huggingface_hub>=0.24.0
 
 # Gradio Interface (REQUIRED)
 gradio>=4.0.0