fd

ProviderCore/registry.py

@@ -156,12 +156,34 @@ def match_provider_name_for_url(url: str) -> Optional[str]:
     try:
         parsed = urlparse(str(url))
         host = (parsed.hostname or "").strip().lower()
+        path = (parsed.path or "").strip()
     except Exception:
         host = ""
+        path = ""
 
     if not host:
         return None
 
+    # Prefer Internet Archive for archive.org links unless the URL clearly refers
+    # to a borrow/loan flow (handled by OpenLibrary provider).
+    #
+    # This keeps direct downloads and item pages routed to `internetarchive`, while
+    # preserving OpenLibrary's scripted borrow pipeline for loan/reader URLs.
+    if host == "openlibrary.org" or host.endswith(".openlibrary.org"):
+        return "openlibrary" if "openlibrary" in _PROVIDERS else None
+
+    if host == "archive.org" or host.endswith(".archive.org"):
+        low_path = str(path or "").lower()
+        is_borrowish = (
+            low_path.startswith("/borrow/")
+            or low_path.startswith("/stream/")
+            or low_path.startswith("/services/loans/")
+            or "/services/loans/" in low_path
+        )
+        if is_borrowish:
+            return "openlibrary" if "openlibrary" in _PROVIDERS else None
+        return "internetarchive" if "internetarchive" in _PROVIDERS else None
+
     for name, provider_class in _PROVIDERS.items():
         domains = getattr(provider_class, "URL_DOMAINS", None)
         if not isinstance(domains, (list, tuple)):