diff --git a/mloader/__main__.py b/mloader/__main__.py
index afcad18..b29d419 100644
--- a/mloader/__main__.py
+++ b/mloader/__main__.py
@@ -105,7 +105,7 @@ def validate_ids(ctx: click.Context, param, value):
     "--raw",
     "-r",
     is_flag=True,
-    default=False,
+    default=True,
     show_default=True,
     help="Save raw images",
     envvar="MLOADER_RAW",
diff --git a/mloader/loader.py b/mloader/loader.py
index 25041d7..fcc9402 100644
--- a/mloader/loader.py
+++ b/mloader/loader.py
@@ -35,13 +35,78 @@ def __init__(
         self.split = split
         self._api_url = "https://jumpg-webapi.tokyo-cdn.com"
         self.session = Session()
+        with open("token.txt", "r", encoding="utf-8") as f:
+            current_token = f.read().strip()
         self.session.headers.update(
             {
                 "User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64; "
-                "rv:72.0) Gecko/20100101 Firefox/72.0"
+                "rv:72.0) Gecko/20100101 Firefox/72.0",
+                "Session-Token": current_token,
             }
         )
 
+    @staticmethod
+    def _parse_api_error(data: bytes) -> Optional[str]:
+        """Extract a human-readable error from a failed API response.
+
+        The API returns errors in protobuf field 2 of Response (not in the
+        'success' field 1 defined in the schema), so they are silently ignored
+        by the generated parser. This method decodes them manually.
+        """
+        def read_varint(buf, pos):
+            result = shift = 0
+            while pos < len(buf):
+                b = buf[pos]; pos += 1
+                result |= (b & 0x7f) << shift
+                shift += 7
+                if not (b & 0x80):
+                    break
+            return result, pos
+
+        def read_ldelim(buf, pos):
+            length, pos = read_varint(buf, pos)
+            return buf[pos:pos + length], pos + length
+
+        try:
+            pos = 0
+            while pos < len(data):
+                tag, pos = read_varint(data, pos)
+                wire_type = tag & 0x7
+                field_num = tag >> 3
+                if wire_type == 2:
+                    chunk, pos = read_ldelim(data, pos)
+                    if field_num == 2:  # failure field
+                        # Scan inner fields for a message containing text
+                        ipos = 0
+                        while ipos < len(chunk):
+                            itag, ipos = read_varint(chunk, ipos)
+                            if itag & 0x7 == 2:
+                                inner, ipos = read_ldelim(chunk, ipos)
+                                # Parse title (field 1) and body (field 2)
+                                title = body = ""
+                                dpos = 0
+                                while dpos < len(inner):
+                                    dtag, dpos = read_varint(inner, dpos)
+                                    if dtag & 0x7 == 2:
+                                        val, dpos = read_ldelim(inner, dpos)
+                                        if dtag >> 3 == 1:
+                                            title = val.decode("utf-8", errors="replace")
+                                        elif dtag >> 3 == 2:
+                                            body = val.decode("utf-8", errors="replace")
+                                    else:
+                                        break
+                                if title:
+                                    return f"{title}: {body}".rstrip(": ") if body else title
+                            else:
+                                break
+                elif wire_type == 0:
+                    _, pos = read_varint(data, pos)
+                else:
+                    break
+        except Exception:
+            pass
+        return None
+
     def _decrypt_image(self, url: str, encryption_hex: str) -> bytearray:
         resp = self.session.get(url)
         data = bytearray(resp.content)
@@ -61,7 +126,13 @@ def _load_pages(self, chapter_id: Union[str, int]) -> MangaViewer:
                 "img_quality": self.quality,
             },
         )
-        return Response.FromString(resp.content).success.manga_viewer
+        viewer = Response.FromString(resp.content).success.manga_viewer
+        if not viewer.pages and not viewer.chapter_id:
+            error = self._parse_api_error(resp.content)
+            raise RuntimeError(
+                error or f"Empty response for chapter {chapter_id}"
+            )
+        return viewer
 
     @lru_cache(None)
     def _get_title_details(self, title_id: Union[str, int]) -> TitleDetailView: