diff --git a/src/chap/backends/llama_cpp.py b/src/chap/backends/llama_cpp.py
index 9eb8f38..ba7b9c7 100644
--- a/src/chap/backends/llama_cpp.py
+++ b/src/chap/backends/llama_cpp.py
@@ -101,5 +101,10 @@ A dialog, where USER interacts with AI. AI is helpful, kind, obedient, honest, a
 
 
 def factory() -> Backend:
-    """Uses the llama.cpp completion web API"""
+    """Uses the llama.cpp completion web API
+
+    Note: Consider using the openai-chatgpt backend with a custom URL instead.
+    The llama.cpp server will automatically apply common chat templates with the
+    openai-chatgpt backend, while chat templates must be manually configured client side
+    with this backend."""
     return LlamaCpp()
diff --git a/src/chap/backends/openai_chatgpt.py b/src/chap/backends/openai_chatgpt.py
index 8f78afc..fd0eb9b 100644
--- a/src/chap/backends/openai_chatgpt.py
+++ b/src/chap/backends/openai_chatgpt.py
@@ -73,7 +73,7 @@ class ChatGPT:
         """The approximate greatest number of tokens to send in a request. When the session is long, the system prompt and 1 or more of the most recent interaction steps are sent."""
 
         url: str = "https://api.openai.com/v1/chat/completions"
-        """The URL of a chatgpt-pcompatible server's completion endpoint."""
+        """The URL of a chatgpt-compatible server's completion endpoint. Notably, llama.cpp's server is compatible with this backend, and can automatically apply common chat templates too."""
 
         temperature: float | None = None
         """The model temperature for sampling"""