diff --git a/modules/openai.py b/modules/openai.py
index 2e0529b..b9b6e21 100644
--- a/modules/openai.py
+++ b/modules/openai.py
@@ -6,6 +6,7 @@ from openai import OpenAI
 
 from nemubot import context
 from nemubot.hooks import hook
+from nemubot.tools import web
 
 from nemubot.module.more import Response
 
@@ -14,18 +15,24 @@ from nemubot.module.more import Response
 
 CLIENT = None
 MODEL = "gpt-4"
+ENDPOINT = None
 
 def load(context):
-    global CLIENT
-    if not context.config or "apikey" not in context.config:
+    global CLIENT, ENDPOINT, MODEL
+    if not context.config or ("apikey" not in context.config and "endpoint" not in context.config):
         raise ImportError ("You need a OpenAI API key in order to use "
                            "this module. Add it to the module configuration: "
                            "\n<module name=\"openai\" "
                            "apikey=\"XXXXXX-XXXXXXXXXX\" endpoint=\"https://...\" model=\"gpt-4\" />")
-    CLIENT = OpenAI(
-        base_url=context.config["endpoint"],
-        api_key=context.config["apikey"],
-    )
+    kwargs = {
+        "api_key": context.config["apikey"] or "",
+    }
+
+    if "endpoint" in context.config:
+        ENDPOINT = context.config["endpoint"]
+        kwargs["base_url"] = ENDPOINT
+
+    CLIENT = OpenAI(**kwargs)
 
     if "model" in context.config:
         MODEL = context.config["model"]
@@ -33,6 +40,32 @@ def load(context):
 
 # MODULE INTERFACE ####################################################
 
+@hook.command("list_models",
+      help="list available LLM")
+def cmd_listllm(msg):
+    llms = web.getJSON(ENDPOINT + "/models", timeout=6)
+    return Response(message=[m for m in map(lambda i: i["id"], llms["data"])], title="Here is the available models", channel=msg.channel)
+
+
+@hook.command("set_model",
+      help="Set the model to use when talking to nemubot")
+def cmd_setllm(msg):
+    if len(msg.args) != 1:
+        raise IMException("Indicate 1 model to use")
+
+    wanted_model = msg.args[0]
+
+    llms = web.getJSON(ENDPOINT + "/models", timeout=6)
+    for model in llms["data"]:
+        if wanted_model == model["id"]:
+            break
+    else:
+        raise IMException("Unable to set such model: unknown")
+
+    MODEL = wanted_model
+    return Response("New model in use: " + wanted_model, channel=msg.channel)
+
+
 @hook.ask()
 def parseask(msg):
     chat_completion = CLIENT.chat.completions.create(