- Change default server host to localhost for improved security.

- Increase default maximum tokens in CLI configuration to 256.
- Refactor and reorganize CLI

crates/inference-engine/src/token_output_stream.rs

@@ -40,7 +40,8 @@ impl TokenOutputStream {
         };
         self.tokens.push(token);
         let text = self.decode(&self.tokens[self.prev_index..])?;
-        if text.len() > prev_text.len() && text.chars().last().unwrap().is_alphanumeric() {
+        if text.len() > prev_text.len() {
+            // Modified to include all tokens, not just alphanumeric ones
             let text = text.split_at(prev_text.len());
             self.prev_index = self.current_index;
             self.current_index = self.tokens.len();