Fix Video Brain URL handling, add API health check, improve error diagnostics

README.md

@@ -19,28 +19,36 @@ An AI-powered search and research assistant with multiple modes:
 - **Agentic**: Multi-agent RAG with planning
 - **Analysis**: Data analysis and insights
 - **Summarize**: Summarize content and documents
+- **Product MVP**: Generate MVP blueprints from ideas
+- **Video Brain**: Analyze YouTube videos
 
 ## Features
 
-- 🔍 Real-time web search with Tavily
+- 🔍 Real-time web search with Tavily (with AI answers)
 - 📚 Document upload and RAG
 - 🤖 LangGraph-powered pipelines
 - 💡 Follow-up question suggestions
 - 🖼️ Image search integration
 - 📊 Knowledge panels
+- 🎥 YouTube video analysis
 
-## Environment Variables
+## Environment Variables (Required!)
 
-Set these secrets in your Hugging Face Space:
+⚠️ **IMPORTANT**: Set these secrets in your Hugging Face Space settings:
 
-- `GROQ_API_KEY`: Your Groq API key
-- `TAVILY_API_KEY`: Your Tavily API key
+1. Go to your Space Settings → Variables and secrets
+2. Add these as **Secrets** (not Repository secrets):
+
+- `GROQ_API_KEY`: Your Groq API key (get from https://console.groq.com)
+- `TAVILY_API_KEY`: Your Tavily API key (get from https://tavily.com)
+
+Without these keys, web search and LLM responses will not work!
 
 ## Tech Stack
 
 - FastAPI backend
 - Streamlit frontend
 - LangChain + LangGraph
-- Groq LLM (llama3-70b-8192)
+- Groq LLM (compound-beta)
 - FAISS vector store
 - Sentence Transformers embeddings

app/api.py

@@ -52,14 +52,25 @@ app.add_middleware(
     allow_credentials=True,
 )
 
+import os
 
 # =======================================================
 # Health Check Endpoint
 # =======================================================
 @app.get("/health")
 async def health_check():
-    """Health check endpoint for container orchestration"""
-    return {"status": "healthy", "service": "perplexity-clone-api"}
+    """Health check endpoint - shows API key status"""
+    groq_key = os.getenv("GROQ_API_KEY", "")
+    tavily_key = os.getenv("TAVILY_API_KEY", "")
+    
+    return {
+        "status": "healthy",
+        "service": "perplexity-clone-api",
+        "api_keys": {
+            "groq": "configured" if groq_key else "MISSING",
+            "tavily": "configured" if tavily_key else "MISSING"
+        }
+    }
 
 
 # =======================================================

streamlit_app.py

@@ -42,6 +42,11 @@ if "product_ideas" not in st.session_state:
 API_URL = os.getenv("BACKEND_URL", "http://127.0.0.1:8000")
 WORKSPACE = "default"
 
+# Check API health on startup
+if "api_checked" not in st.session_state:
+    st.session_state.api_checked = False
+    st.session_state.api_status = None
+
 # MODE MAPPING - All 8 modes with correct backend endpoints
 MODES = {
     "Automatic": {
@@ -456,6 +461,37 @@ def get_css():
 st.markdown(get_css(), unsafe_allow_html=True)
 
 
+# =====================================
+# API HEALTH CHECK
+# =====================================
+def check_api_health():
+    """Check if API is healthy and keys are configured."""
+    try:
+        resp = requests.get(f"{API_URL}/health", timeout=5)
+        if resp.ok:
+            return resp.json()
+    except:
+        pass
+    return None
+
+# Check API on first load
+if not st.session_state.api_checked:
+    st.session_state.api_status = check_api_health()
+    st.session_state.api_checked = True
+
+# Show warnings if API keys are missing
+if st.session_state.api_status:
+    api_keys = st.session_state.api_status.get("api_keys", {})
+    missing_keys = []
+    if api_keys.get("groq") == "MISSING":
+        missing_keys.append("GROQ_API_KEY")
+    if api_keys.get("tavily") == "MISSING":
+        missing_keys.append("TAVILY_API_KEY")
+    
+    if missing_keys:
+        st.error(f"⚠️ Missing API keys: {', '.join(missing_keys)}. Please configure them in your Hugging Face Space settings under 'Variables and secrets'.")
+
+
 # =====================================
 # HELPER FUNCTIONS
 # =====================================
@@ -635,11 +671,16 @@ if st.session_state.mode == "Video Brain" and not st.session_state.current_resul
     <div style="text-align: center; padding: 20px; margin: 20px auto; max-width: 700px; 
                 background: linear-gradient(135deg, #3B82F6 0%, #8B5CF6 100%); 
                 border-radius: 16px; color: white;">
-        <h3 style="margin: 0; font-size: 24px;">🎥 Video Brain – Understand Any YouTube Lecture</h3>
-        <p style="margin: 10px 0 0; opacity: 0.9;">🔵 Paste YouTube URL below, then ask questions</p>
+        <h3 style="margin: 0; font-size: 24px;">🎥 Video Brain – Understand Any YouTube Video</h3>
+        <p style="margin: 10px 0 0; opacity: 0.9;">🔵 Two ways to use:</p>
     </div>
     """, unsafe_allow_html=True)
     
+    st.markdown("""
+    **Option 1:** Paste a YouTube URL directly in the search box and click Search  
+    **Option 2:** Load the URL below first, then ask questions
+    """)
+    
     # YouTube URL input - auto-loads on change
     youtube_url = st.text_input(
         "YouTube URL",
@@ -689,26 +730,43 @@ if st.session_state.uploaded_files:
 # =====================================
 if submit and query.strip():
     extra_data = None
+    actual_query = query.strip()
     
     # For Video Brain mode, include the YouTube URL
     if st.session_state.mode == "Video Brain":
-        if st.session_state.video_loaded and st.session_state.youtube_url:
+        # Check if query itself contains a YouTube URL
+        query_text = query.strip()
+        url_in_query = None
+        for word in query_text.split():
+            if "youtube.com" in word or "youtu.be" in word:
+                url_in_query = word
+                break
+        
+        if url_in_query:
+            # User pasted URL directly in the query
+            st.session_state.youtube_url = url_in_query
+            st.session_state.video_loaded = True
+            extra_data = {"youtube_url": url_in_query}
+            # If query is JUST the URL, ask a default question
+            if query_text == url_in_query:
+                actual_query = "Summarize this video and give me the key takeaways"
+        elif st.session_state.video_loaded and st.session_state.youtube_url:
             extra_data = {"youtube_url": st.session_state.youtube_url}
         else:
-            st.warning("⚠️ Please load a YouTube video first!")
+            st.warning("⚠️ Please paste a YouTube URL in the search box or load one above!")
             st.stop()
     
     # For Product MVP mode, save to ideas history
     if st.session_state.mode == "Product MVP":
         st.session_state.product_ideas.append({
-            "idea": query.strip(),
+            "idea": actual_query,
             "time": "just now"
         })
     
     with st.spinner(f"🔄 {st.session_state.mode}..."):
-        result = call_api(query.strip(), st.session_state.mode, extra_data)
+        result = call_api(actual_query, st.session_state.mode, extra_data)
         st.session_state.current_result = {
-            "query": query.strip(),
+            "query": actual_query,
             "mode": st.session_state.mode,
             "data": result
         }