diff --git a/src/backend/ggml.rs b/src/backend/ggml.rs
index 2ea41ad..420e85a 100644
--- a/src/backend/ggml.rs
+++ b/src/backend/ggml.rs
@@ -1,6 +1,4 @@
-use crate::{
-    error, utils::gen_chat_id, ServerInfo, GLOBAL_RAG_PROMPT, MULTI_RETRIEVAL, SERVER_INFO,
-};
+use crate::{error, utils::gen_chat_id, ServerInfo, GLOBAL_RAG_PROMPT, SERVER_INFO};
 use chat_prompts::{error as ChatPromptsError, MergeRagContext, MergeRagContextPolicy};
 use endpoints::{
     chat::{ChatCompletionRequest, ChatCompletionRequestMessage, ChatCompletionUserMessageContent},
@@ -486,7 +484,7 @@ async fn retrieve_context(
             }
 
             // join the user messages in the context window into a single string
-            let query_text = if last_messages.len() > 0 {
+            let query_text = if !last_messages.is_empty() {
                 info!(target: "stdout", "Found the latest {} user messages.", last_messages.len());
 
                 last_messages.reverse();
diff --git a/src/main.rs b/src/main.rs
index 2673121..75612c0 100644
--- a/src/main.rs
+++ b/src/main.rs
@@ -29,8 +29,6 @@ type Error = Box<dyn std::error::Error + Send + Sync + 'static>;
 pub(crate) static GLOBAL_RAG_PROMPT: OnceCell<String> = OnceCell::new();
 // server info
 pub(crate) static SERVER_INFO: OnceCell<ServerInfo> = OnceCell::new();
-// allow multi-retrieval
-pub(crate) static MULTI_RETRIEVAL: OnceCell<bool> = OnceCell::new();
 
 // default port
 const DEFAULT_PORT: &str = "8080";
@@ -115,9 +113,6 @@ struct Cli {
     /// Maximum number of tokens each chunk contains
     #[arg(long, default_value = "100", value_parser = clap::value_parser!(usize))]
     chunk_capacity: usize,
-    /// Allow multi-retrieval
-    #[arg(long, default_value = "true")]
-    multi_retrieval: bool,
     /// Socket address of LlamaEdge-RAG API Server instance. For example, `0.0.0.0:8080`.
     #[arg(long, default_value = None, value_parser = clap::value_parser!(SocketAddr), group = "socket_address_group")]
     socket_addr: Option<SocketAddr>,
@@ -300,12 +295,6 @@ async fn main() -> Result<(), ServerError> {
     // log chunk capacity
     info!(target: "stdout", "chunk_capacity: {}", &cli.chunk_capacity);
 
-    // log multi-retrieval
-    info!(target: "stdout", "multi_retrieval: {}", &cli.multi_retrieval);
-    MULTI_RETRIEVAL
-        .set(cli.multi_retrieval)
-        .map_err(|_| ServerError::Operation("Failed to set `MULTI_RETRIEVAL`.".to_string()))?;
-
     // RAG policy
     info!(target: "stdout", "rag_policy: {}", &cli.policy);