diff --git a/kai/kai_config.py b/kai/kai_config.py
index 04019392..62d81c86 100644
--- a/kai/kai_config.py
+++ b/kai/kai_config.py
@@ -157,6 +157,7 @@ class SupportedModelProviders(StrEnum):
     CHAT_GOOGLE_GENERATIVE_AI = "ChatGoogleGenerativeAI"
     AZURE_CHAT_OPENAI = "AzureChatOpenAI"
     CHAT_DEEP_SEEK = "ChatDeepSeek"
+    VLLM_OPENAI = "VLLMOpenAI"
 
 
 class KaiConfigModels(BaseModel):
diff --git a/kai/llm_interfacing/model_provider.py b/kai/llm_interfacing/model_provider.py
index d4abe499..376ca23f 100644
--- a/kai/llm_interfacing/model_provider.py
+++ b/kai/llm_interfacing/model_provider.py
@@ -1,7 +1,7 @@
 from __future__ import annotations
 
 import os
-from typing import Any, Optional
+from typing import Any, Optional, assert_never
 
 from langchain_aws import ChatBedrock
 from langchain_community.chat_models.fake import FakeListChatModel
@@ -17,7 +17,7 @@
 from pydantic.v1.utils import deep_update
 
 from kai.cache import Cache, CachePathResolver, SimplePathResolver
-from kai.kai_config import KaiConfigModels
+from kai.kai_config import KaiConfigModels, SupportedModelProviders
 from kai.logging.logging import get_logger
 
 LOG = get_logger(__name__)
@@ -43,7 +43,7 @@ def __init__(
 
         # Set the model class, model args, and model id based on the provider
         match config.provider:
-            case "ChatOllama":
+            case SupportedModelProviders.CHAT_OLLAMA:
                 model_class = ChatOllama
 
                 defaults = {
@@ -56,7 +56,7 @@ def __init__(
                 model_args = deep_update(defaults, config.args)
                 model_id = model_args["model"]
 
-            case "ChatOpenAI":
+            case SupportedModelProviders.CHAT_OPENAI:
                 model_class = ChatOpenAI
 
                 defaults = {
@@ -96,7 +96,7 @@ def _get_request_payload(
                     if "temperature" in model_args:
                         del model_args["temperature"]
 
-            case "ChatBedrock":
+            case SupportedModelProviders.CHAT_BEDROCK:
                 model_class = ChatBedrock
 
                 defaults = {
@@ -106,7 +106,7 @@ def _get_request_payload(
                 model_args = deep_update(defaults, config.args)
                 model_id = model_args["model_id"]
 
-            case "FakeListChatModel":
+            case SupportedModelProviders.FAKE_LIST_CHAT_MODEL:
                 model_class = FakeListChatModel
 
                 defaults = {
@@ -132,7 +132,7 @@ def _get_request_payload(
                 model_args = deep_update(defaults, config.args)
                 model_id = "fake-list-chat-model"
 
-            case "ChatGoogleGenerativeAI":
+            case SupportedModelProviders.CHAT_GOOGLE_GENERATIVE_AI:
                 model_class = ChatGoogleGenerativeAI
                 api_key = os.getenv("GOOGLE_API_KEY", "dummy_value")
                 defaults = {
@@ -144,7 +144,7 @@ def _get_request_payload(
                 model_args = deep_update(defaults, config.args)
                 model_id = model_args["model"]
 
-            case "AzureChatOpenAI":
+            case SupportedModelProviders.AZURE_CHAT_OPENAI:
                 model_class = AzureChatOpenAI
 
                 defaults = {
@@ -159,7 +159,7 @@ def _get_request_payload(
                 model_args = deep_update(defaults, config.args)
                 model_id = model_args["azure_deployment"]
 
-            case "ChatDeepSeek":
+            case SupportedModelProviders.CHAT_DEEP_SEEK:
                 model_class = ChatDeepSeek
 
                 defaults = {
@@ -173,7 +173,7 @@ def _get_request_payload(
                 model_args = deep_update(defaults, config.args)
                 model_id = model_args["model"]
 
-            case "VLLMOpenAI":
+            case SupportedModelProviders.VLLM_OPENAI:
                 model_class = VLLMOpenAI
 
                 defaults = {}
@@ -182,6 +182,7 @@ def _get_request_payload(
                 model_id = model_args["model_name"]
 
             case _:
+                assert_never(config.provider)
                 raise Exception(f"Unrecognized provider '{config.provider}'")
 
         self.provider_id: str = config.provider
@@ -278,7 +279,7 @@ def invoke(
 
     def response_to_base_message(self, llm_response: Any) -> BaseMessage:
         if isinstance(llm_response, str):
-            return BaseMessage(content=llm_response)
+            return BaseMessage(content=llm_response, type="ai")
         elif isinstance(llm_response, BaseMessage):
             return llm_response