diff --git a/config/settings.py b/config/settings.py
index e7c0c1a..8c5808d 100644
--- a/config/settings.py
+++ b/config/settings.py
@@ -50,6 +50,8 @@ def load_config():
             os.getenv("CLOVER_MAX_HISTORY_MESSAGES", "20")
         ),
         "max_history_tokens": int(os.getenv("CLOVER_MAX_HISTORY_TOKENS", "8000")),
+        "max_retries": int(os.getenv("CLOVER_MAX_RETRIES", "3")),
+        "retry_base_delay": float(os.getenv("CLOVER_RETRY_BASE_DELAY", "2")),
     }
 
     # Debug output if enabled
diff --git a/models/api_client.py b/models/api_client.py
index b3f7d31..fc7fb80 100644
--- a/models/api_client.py
+++ b/models/api_client.py
@@ -5,6 +5,7 @@ Handles communication with the OpenAI-compatible LLM server
 
 import json
 import os
+import time
 from typing import Any, Dict, Optional
 
 import requests
@@ -55,7 +56,10 @@ class APIClient:
         self, endpoint: str, method: str = "GET", data: Optional[Dict] = None
     ) -> Dict[str, Any]:
         """
-        Make a request to the LLM API server
+        Make a request to the LLM API server with retry and exponential backoff.
+
+        Retries up to 3 times with exponential backoff (2s base delay) for
+        transient failures (5xx errors, connection errors, timeouts).
 
         Uses requests.Session with explicit SSL verification control.
 
@@ -67,33 +71,74 @@ class APIClient:
         Returns:
             dict: Response from the API
         """
-        try:
-            # Ensure base_url doesn't have trailing slash and endpoint has leading slash
-            base = self.base_url.rstrip("/")
-            endpoint = endpoint if endpoint.startswith("/") else f"/{endpoint}"
-            url = f"{base}{endpoint}"
+        max_retries = self.config.get("max_retries", 3)
+        base_delay = self.config.get("retry_base_delay", 2)
 
-            kwargs = {
-                "timeout": self.config.get("timeout", 300),
-                "verify": self.ssl_verify,
+        last_exception = None
+
+        for attempt in range(max_retries + 1):
+            try:
+                kwargs = {
+                    "timeout": self.config.get("timeout", 300),
+                    "verify": self.ssl_verify,
+                }
+
+                if method == "GET":
+                    response = self.session.get(url, **kwargs)
+                elif method == "POST":
+                    response = self.session.post(url, json=data, **kwargs)
+                else:
+                    raise ValueError(f"Unsupported HTTP method: {method}")
+
+                # Retry on 5xx server errors
+                if response.status_code >= 500 and attempt < max_retries:
+                    delay = base_delay * (2**attempt)
+                    print(
+                        f"Server error {response.status_code}, "
+                        f"retrying in {delay}s (attempt {attempt + 1}/{max_retries})"
+                    )
+                    time.sleep(delay)
+                    continue
+
+                response.raise_for_status()
+                return {"success": True, "data": response.json()}
+
+            except requests.exceptions.Timeout as e:
+                last_exception = e
+                if attempt < max_retries:
+                    delay = base_delay * (2**attempt)
+                    print(
+                        f"Timeout, retrying in {delay}s "
+                        f"(attempt {attempt + 1}/{max_retries})"
+                    )
+                    time.sleep(delay)
+                    continue
+
+            except requests.exceptions.ConnectionError as e:
+                last_exception = e
+                if attempt < max_retries:
+                    delay = base_delay * (2**attempt)
+                    print(
+                        f"Connection error, retrying in {delay}s "
+                        f"(attempt {attempt + 1}/{max_retries})"
+                    )
+                    time.sleep(delay)
+                    continue
+
+            except requests.exceptions.RequestException as e:
+                return {"success": False, "error": f"HTTP Request failed: {str(e)}"}
+            except json.JSONDecodeError as e:
+                return {"success": False, "error": f"Invalid JSON response: {str(e)}"}
+            except Exception as e:
+                return {"success": False, "error": f"Unexpected error: {str(e)}"}
+
+        # All retries exhausted
+        if last_exception:
+            return {
+                "success": False,
+                "error": f"Request failed after {max_retries} retries: {str(last_exception)}",
             }
-
-            if method == "GET":
-                response = self.session.get(url, **kwargs)
-            elif method == "POST":
-                response = self.session.post(url, json=data, **kwargs)
-            else:
-                raise ValueError(f"Unsupported HTTP method: {method}")
-
-            response.raise_for_status()
-            return {"success": True, "data": response.json()}
-
-        except requests.exceptions.RequestException as e:
-            return {"success": False, "error": f"HTTP Request failed: {str(e)}"}
-        except json.JSONDecodeError as e:
-            return {"success": False, "error": f"Invalid JSON response: {str(e)}"}
-        except Exception as e:
-            return {"success": False, "error": f"Unexpected error: {str(e)}"}
+        return {"success": False, "error": "Request failed after maximum retries"}
 
     def list_models(self) -> Dict[str, Any]:
         """