diff --git a/gallery_dl/extractor/patreon.py b/gallery_dl/extractor/patreon.py
index 60549983..fefef48c 100644
--- a/gallery_dl/extractor/patreon.py
+++ b/gallery_dl/extractor/patreon.py
@@ -69,6 +69,7 @@ class PatreonExtractor(Extractor):
                     post["type"] = "content"
                     yield Message.Url, url, text.nameext_from_url(url, post)
 
+            # Metadata for post using dummy url for formatting
             post.update({"metadata_only": True})
             url = post.get("creator").get("image_url")
             yield Message.Metadata, url, text.nameext_from_url(url, post)
diff --git a/gallery_dl/job.py b/gallery_dl/job.py
index 19878db1..972174a4 100644
--- a/gallery_dl/job.py
+++ b/gallery_dl/job.py
@@ -99,9 +99,8 @@ class Job():
 
         elif msg[0] == Message.Metadata:
             _, url, kwds = msg
-            if self.pred_url(url, kwds):
-                self.update_kwdict(kwds)
-                self.handle_url(url, kwds)
+            self.update_kwdict(kwds)
+            self.handle_metadata(url, kwds)
 
         elif msg[0] == Message.Version:
             if msg[1] != 1:
@@ -194,9 +193,6 @@ class DownloadJob(Job):
             for pp in postprocessors:
                 pp.prepare(pathfmt)
 
-        if kwdict.get("metadata_only"):
-            return
-
         if pathfmt.exists(archive):
             self.handle_skip()
             return
@@ -238,6 +234,19 @@ class DownloadJob(Job):
                 pp.run_after(pathfmt)
         self._skipcnt = 0
 
+    def handle_metadata(self, url, kwdict, fallback=None):
+        """Download the resource specified in 'url'"""
+        postprocessors = self.postprocessors
+        pathfmt = self.pathfmt
+
+        # prepare download
+        pathfmt.set_filename(kwdict)
+
+        if postprocessors:
+            for pp in postprocessors:
+                pp.prepare(pathfmt)
+        return
+
     def handle_urllist(self, urls, kwdict):
         """Download the resource specified in 'url'"""
         fallback = iter(urls)
@@ -510,6 +519,9 @@ class DataJob(Job):
     def handle_url(self, url, kwdict):
         self.data.append((Message.Url, url, self.filter(kwdict)))
 
+    def handle_metadata(self, url, kwdict):
+        self.data.append((Message.Url, url, self.filter(kwdict)))
+
     def handle_urllist(self, urls, kwdict):
         self.data.append((Message.Urllist, list(urls), self.filter(kwdict)))
 
diff --git a/gallery_dl/postprocessor/metadata_bypost.py b/gallery_dl/postprocessor/metadata_bypost.py
index 05781742..cc5dc165 100644
--- a/gallery_dl/postprocessor/metadata_bypost.py
+++ b/gallery_dl/postprocessor/metadata_bypost.py
@@ -17,6 +17,7 @@ class Metadata_bypostPP(MetadataPP):
         MetadataPP.__init__(self, pathfmt, options)
 
     def prepare(self, pathfmt):
+        # Only run this processor on metadata messages, not individual images.
         if pathfmt.kwdict.get("metadata_only"):
             MetadataPP.run(self, pathfmt)