naive_hook.diff

diff --git a/autogen/hooks_utils.py b/autogen/hooks_utils.py
new file mode 100644
index 00000000..7de0e876
--- /dev/null
+++ b/autogen/hooks_utils.py
@@ -0,0 +1,24 @@
+from typing import Callable
+from collections import defaultdict
+
+class Hooks:
+    _hooks = defaultdict(list)
+
+    def register(self, name: str, func: Callable) -> None:
+        self._hooks[name].append(func)
+
+    def trigger(self, name: str, *args, **kwargs) -> None:
+        for func in self._hooks[name]:
+            func(*args, **kwargs)
+
+    @classmethod
+    def instance(cls):
+        if not hasattr(cls, "_instance"):
+            cls._instance = cls()
+        return cls._instance
+
+if __name__ == "__main__":
+    hooks = Hooks.instance()
+    hooks.register("test", lambda x: print('1', x))
+    hooks.register("test", lambda x: print('2', x))
+    hooks.trigger('test', "p1")
diff --git a/autogen/oai/client.py b/autogen/oai/client.py
index 1bdfd835..e191b174 100644
--- a/autogen/oai/client.py
+++ b/autogen/oai/client.py
@@ -15,6 +15,8 @@ from autogen.oai.openai_utils import get_key, OAI_PRICE1K
 from autogen.token_count_utils import count_token
 from autogen._pydantic import model_dump
 
+from autogen.hooks_utils import Hooks
+
 TOOL_ENABLED = False
 try:
     import openai
@@ -287,6 +289,8 @@ class OpenAIWrapper:
             else:
                 # add cost calculation before caching no matter filter is passed or not
                 response.cost = self.cost(response)
+                Hooks.instance().trigger("after_llm", input=full_config, output=response)
+
                 self._update_usage_summary(response, use_cache=False)
                 if cache_seed is not None:
                     # Cache the response
diff --git a/test/test_hooks.py b/test/test_hooks.py
new file mode 100644
index 00000000..01e9c577
--- /dev/null
+++ b/test/test_hooks.py
@@ -0,0 +1,16 @@
+from autogen import config_list_from_json, OpenAIWrapper
+from autogen.hooks_utils import Hooks
+import json
+
+config_list = config_list_from_json(env_or_file="OAI_CONFIG_LIST")
+
+client = OpenAIWrapper()
+
+def show_me(input, output):
+    print('input', input)
+    print('output', output)
+    print('Cost', output.cost)
+
+hooks = Hooks.instance().register('after_llm', show_me)
+
+client.create(messages=[{"role": "system", "content": "you are a cat"}, {"role": "user", "content": "2+1="}], model='gpt-3.5-turbo', cache_seed=None)
\ No newline at end of file