diff --git a/backend/main.py b/backend/main.py
index 9021161..3f12431 100644
--- a/backend/main.py
+++ b/backend/main.py
@@ -2,15 +2,16 @@
 import os
 import time
 import asyncio
+import json
 from typing import Optional, Dict, Any
 from datetime import datetime
 
 from dotenv import load_dotenv
 from fastapi import FastAPI, HTTPException
 from fastapi.middleware.cors import CORSMiddleware
+from fastapi.responses import StreamingResponse
 from pydantic import BaseModel, Field
 import uvicorn
-import json
 
 from agents import NeuroscienceAssistant
 
@@ -54,9 +55,8 @@ class ChatResponse(BaseModel):
 # Lightweight health helpers
 
 def _vector_check_sync() -> bool:
-    
     try:
-        from retrieval import Retriever  # local import to avoid import penalty on startup
+        from retrieval import Retriever
         r = Retriever()
         return bool(getattr(r, "is_enabled", False))
     except Exception:
@@ -109,6 +109,57 @@ async def health():
     }
 
 
+@app.get("/api/chat/stream", tags=["Chat"])
+async def chat_stream_endpoint(query: str, session_id: str = "default"):
+    """
+    Stream chat responses using Server-Sent Events (SSE).
+    Tokens appear in real-time as the LLM generates them.
+    
+    This provides a better user experience by showing responses
+    as they are generated, similar to ChatGPT.
+    """
+    async def generate():
+        try:
+            # Send initial connection message
+            yield f"data: {json.dumps({'type': 'start', 'message': 'Connected'})}\n\n"
+            
+            # Get the response from the assistant
+            response_text = await assistant.handle_chat(
+                session_id=session_id,
+                query=query,
+                reset=False,
+            )
+            
+            # Stream the response in chunks (word-by-word simulation)
+            words = response_text.split(' ')
+            chunk_size = 3  # Send 3 words at a time for smooth streaming
+            
+            for i in range(0, len(words), chunk_size):
+                chunk = ' '.join(words[i:i + chunk_size])
+                if i > 0:
+                    chunk = ' ' + chunk
+                yield f"data: {json.dumps({'type': 'token', 'content': chunk})}\n\n"
+                await asyncio.sleep(0.03)  # 30ms delay for streaming effect
+            
+            # Send completion message
+            yield f"data: {json.dumps({'type': 'done', 'message': 'Complete'})}\n\n"
+            
+        except asyncio.TimeoutError:
+            yield f"data: {json.dumps({'type': 'error', 'message': 'Request timed out. Please try a simpler query.'})}\n\n"
+        except Exception as e:
+            yield f"data: {json.dumps({'type': 'error', 'message': str(e)})}\n\n"
+    
+    return StreamingResponse(
+        generate(),
+        media_type="text/event-stream",
+        headers={
+            "Cache-Control": "no-cache",
+            "Connection": "keep-alive",
+            "X-Accel-Buffering": "no",
+        }
+    )
+
+
 @app.post("/api/chat", response_model=ChatResponse, tags=["Chat"])
 async def chat_endpoint(msg: ChatMessage):
     try:
@@ -138,8 +189,6 @@ async def chat_endpoint(msg: ChatMessage):
         )
 
 
-
-
 @app.post("/api/session/reset", tags=["Chat"])
 async def reset_session(payload: Dict[str, str]):
     sid = (payload or {}).get("session_id") or "default"
@@ -154,7 +203,7 @@ async def reset_session(payload: Dict[str, str]):
         "main:app",
         host=os.getenv("HOST", "0.0.0.0"),
         port=int(os.getenv("PORT", "8000")),
-        reload=True, 
+        reload=True,
         log_level="info",
         proxy_headers=True,
-    )
+    )
\ No newline at end of file