diff --git a/reginald/cli.py b/reginald/cli.py
index 0b9c79d..c2f4ab5 100644
--- a/reginald/cli.py
+++ b/reginald/cli.py
@@ -32,6 +32,8 @@
     "streaming": "Whether to use streaming for the chat interaction.",
     "slack_app_token": "Slack app token for the bot.",
     "slack_bot_token": "Slack bot token for the bot.",
+    "host": "Host to listen on.",
+    "port": "Port to listen on.",
 }
 
 cli = typer.Typer(context_settings={"help_option_names": ["-h", "--help"]})
@@ -261,6 +263,12 @@ def app(
     device: Annotated[
         str, typer.Option(envvar="LLAMA_INDEX_DEVICE", help=HELP_TEXT["device"])
     ] = DEFAULT_ARGS["device"],
+    host: Annotated[
+        str, typer.Option(envvar="REGINALD_HOST", help=HELP_TEXT["host"])
+    ] = DEFAULT_ARGS["host"],
+    port: Annotated[
+        int, typer.Option(envvar="REGINALD_PORT", help=HELP_TEXT["port"])
+    ] = DEFAULT_ARGS["port"],
 ) -> None:
     """
     Sets up the response model and then creates a
@@ -273,6 +281,8 @@ def app(
     set_up_logging_config(level=20)
     main(
         cli="app",
+        host=host,
+        port=port,
         model=model,
         model_name=model_name,
         mode=mode,
diff --git a/reginald/defaults.py b/reginald/defaults.py
index 7aa705a..413bcd5 100644
--- a/reginald/defaults.py
+++ b/reginald/defaults.py
@@ -19,4 +19,6 @@
     "is_path": False,
     "n_gpu_layers": 0,
     "device": "auto",
+    "host": "0.0.0.0",
+    "port": 8000,
 }
diff --git a/reginald/models/app.py b/reginald/models/app.py
index beedae0..5d1c1a2 100644
--- a/reginald/models/app.py
+++ b/reginald/models/app.py
@@ -2,6 +2,7 @@
 from fastapi import FastAPI
 from pydantic import BaseModel
 
+from reginald.defaults import DEFAULT_ARGS
 from reginald.models.setup_llm import setup_llm
 
 
@@ -43,8 +44,17 @@ async def channel_mention(query: Query):
     return app
 
 
-def run_reginald_app(**kwargs) -> None:
+def run_reginald_app(
+    host: str | None = DEFAULT_ARGS["host"],
+    port: int | None = DEFAULT_ARGS["port"],
+    **kwargs
+) -> None:
+    if host is None:
+        host = DEFAULT_ARGS["host"]
+    if port is None:
+        port = DEFAULT_ARGS["port"]
+
     # set up response model
     response_model = setup_llm(**kwargs)
     app: FastAPI = create_reginald_app(response_model)
-    uvicorn.run(app, host="0.0.0.0", port=8000)
+    uvicorn.run(app, host=host, port=port)
diff --git a/reginald/run.py b/reginald/run.py
index 75a3cd6..26728c2 100644
--- a/reginald/run.py
+++ b/reginald/run.py
@@ -11,6 +11,8 @@ def main(
     which_index: str | None = None,
     slack_app_token: str | None = None,
     slack_bot_token: str | None = None,
+    host: str | None = None,
+    port: int | None = None,
     **kwargs,
 ):
     # initialise logging
@@ -40,7 +42,9 @@ def main(
     elif cli == "app":
         from reginald.models.app import run_reginald_app
 
-        run_reginald_app(data_dir=data_dir, which_index=which_index, **kwargs)
+        run_reginald_app(
+            host=host, port=port, data_dir=data_dir, which_index=which_index, **kwargs
+        )
     elif cli == "chat":
         import warnings