diff --git a/examples/lightrag_openai_compatible_stream_demo.py b/examples/lightrag_openai_compatible_stream_demo.py
new file mode 100644
index 00000000..9345ada5
--- /dev/null
+++ b/examples/lightrag_openai_compatible_stream_demo.py
@@ -0,0 +1,55 @@
+import os
+import inspect
+from lightrag import LightRAG
+from lightrag.llm import openai_complete, openai_embedding
+from lightrag.utils import EmbeddingFunc
+from lightrag.lightrag import always_get_an_event_loop
+from lightrag import QueryParam
+
+# WorkingDir
+ROOT_DIR = os.path.dirname(os.path.abspath(__file__))
+WORKING_DIR = os.path.join(ROOT_DIR, "dickens")
+if not os.path.exists(WORKING_DIR):
+    os.mkdir(WORKING_DIR)
+print(f"WorkingDir: {WORKING_DIR}")
+
+api_key = "empty"
+rag = LightRAG(
+    working_dir=WORKING_DIR,
+    llm_model_func=openai_complete,
+    llm_model_name="qwen2.5-14b-instruct@4bit",
+    llm_model_max_async=4,
+    llm_model_max_token_size=32768,
+    llm_model_kwargs={"base_url": "http://127.0.0.1:1234/v1", "api_key": api_key},
+    embedding_func=EmbeddingFunc(
+        embedding_dim=1024,
+        max_token_size=8192,
+        func=lambda texts: openai_embedding(
+            texts=texts,
+            model="text-embedding-bge-m3",
+            base_url="http://127.0.0.1:1234/v1",
+            api_key=api_key,
+        ),
+    ),
+)
+
+with open("./book.txt", "r", encoding="utf-8") as f:
+    rag.insert(f.read())
+
+resp = rag.query(
+    "What are the top themes in this story?",
+    param=QueryParam(mode="hybrid", stream=True),
+)
+
+
+async def print_stream(stream):
+    async for chunk in stream:
+        if chunk:
+            print(chunk, end="", flush=True)
+
+
+loop = always_get_an_event_loop()
+if inspect.isasyncgen(resp):
+    loop.run_until_complete(print_stream(resp))
+else:
+    print(resp)
diff --git a/lightrag/llm.py b/lightrag/llm.py
index d02d5350..05f4bf00 100644
--- a/lightrag/llm.py
+++ b/lightrag/llm.py
@@ -91,26 +91,40 @@ async def openai_complete_if_cache(
         response = await openai_async_client.chat.completions.create(
             model=model, messages=messages, **kwargs
         )
-    content = response.choices[0].message.content
-    if r"\u" in content:
-        content = content.encode("utf-8").decode("unicode_escape")
 
-    # Save to cache
-    await save_to_cache(
-        hashing_kv,
-        CacheData(
-            args_hash=args_hash,
-            content=content,
-            model=model,
-            prompt=prompt,
-            quantized=quantized,
-            min_val=min_val,
-            max_val=max_val,
-            mode=mode,
-        ),
-    )
+    if hasattr(response, "__aiter__"):
 
-    return content
+        async def inner():
+            async for chunk in response:
+                content = chunk.choices[0].delta.content
+                if content is None:
+                    continue
+                if r"\u" in content:
+                    content = content.encode("utf-8").decode("unicode_escape")
+                yield content
+
+        return inner()
+    else:
+        content = response.choices[0].message.content
+        if r"\u" in content:
+            content = content.encode("utf-8").decode("unicode_escape")
+
+        # Save to cache
+        await save_to_cache(
+            hashing_kv,
+            CacheData(
+                args_hash=args_hash,
+                content=content,
+                model=model,
+                prompt=prompt,
+                quantized=quantized,
+                min_val=min_val,
+                max_val=max_val,
+                mode=mode,
+            ),
+        )
+
+        return content
 
 
 @retry(
@@ -431,7 +445,7 @@ async def ollama_model_if_cache(
 
     response = await ollama_client.chat(model=model, messages=messages, **kwargs)
     if stream:
-        """ cannot cache stream response """
+        """cannot cache stream response"""
 
         async def inner():
             async for chunk in response:
@@ -613,6 +627,22 @@ class GPTKeywordExtractionFormat(BaseModel):
     low_level_keywords: List[str]
 
 
+async def openai_complete(
+    prompt, system_prompt=None, history_messages=[], keyword_extraction=False, **kwargs
+) -> Union[str, AsyncIterator[str]]:
+    keyword_extraction = kwargs.pop("keyword_extraction", None)
+    if keyword_extraction:
+        kwargs["response_format"] = "json"
+    model_name = kwargs["hashing_kv"].global_config["llm_model_name"]
+    return await openai_complete_if_cache(
+        model_name,
+        prompt,
+        system_prompt=system_prompt,
+        history_messages=history_messages,
+        **kwargs,
+    )
+
+
 async def gpt_4o_complete(
     prompt, system_prompt=None, history_messages=[], keyword_extraction=False, **kwargs
 ) -> str:
@@ -1089,12 +1119,14 @@ async def save_to_cache(hashing_kv, cache_data: CacheData):
     mode_cache[cache_data.args_hash] = {
         "return": cache_data.content,
         "model": cache_data.model,
-        "embedding": cache_data.quantized.tobytes().hex()
-        if cache_data.quantized is not None
-        else None,
-        "embedding_shape": cache_data.quantized.shape
-        if cache_data.quantized is not None
-        else None,
+        "embedding": (
+            cache_data.quantized.tobytes().hex()
+            if cache_data.quantized is not None
+            else None
+        ),
+        "embedding_shape": (
+            cache_data.quantized.shape if cache_data.quantized is not None else None
+        ),
         "embedding_min": cache_data.min_val,
         "embedding_max": cache_data.max_val,
         "original_prompt": cache_data.prompt,