diff --git a/lightrag/llm/openai.py b/lightrag/llm/openai.py
index d367abc7..07942101 100644
--- a/lightrag/llm/openai.py
+++ b/lightrag/llm/openai.py
@@ -10,8 +10,18 @@ import pipmaster as pm
 if not pm.is_installed("openai"):
     pm.install("openai")
 
+# Try to import Langfuse for LLM observability (optional)
+# Falls back to standard OpenAI client if not available
+try:
+    from langfuse.openai import AsyncOpenAI
+    LANGFUSE_ENABLED = True
+    logger.info("Langfuse observability enabled for OpenAI client")
+except ImportError:
+    from openai import AsyncOpenAI
+    LANGFUSE_ENABLED = False
+    logger.debug("Langfuse not available, using standard OpenAI client")
+
 from openai import (
-    AsyncOpenAI,
     APIConnectionError,
     RateLimitError,
     APITimeoutError,
diff --git a/pyproject.toml b/pyproject.toml
index 57e1b765..8eabb37c 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -113,6 +113,11 @@ offline = [
     "lightrag-hku[offline-docs,offline-storage,offline-llm]",
 ]
 
+observability = [
+    # LLM observability and tracing dependencies
+    "langfuse>=3.8.1",
+]
+
 [project.scripts]
 lightrag-server = "lightrag.api.lightrag_server:main"
 lightrag-gunicorn = "lightrag.api.run_with_gunicorn:main"