diff --git a/lib/jsduck/app.rb b/lib/jsduck/app.rb
index c80c1771d3dcbde62234b0ab4fedae1c6462adc2..15599d9748e22459362cb24d25455f5fc877ebb1 100644
--- a/lib/jsduck/app.rb
+++ b/lib/jsduck/app.rb
@@ -47,7 +47,7 @@ module JsDuck
       result = aggregate(parsed_files)
       @relations = filter_classes(result)
       InheritDoc.new(@relations).resolve_all
-      Importer.import(@opts.imports, @relations)
+      Importer.import(@opts.imports, @relations, @parallel)
       Lint.new(@relations).run
 
       # Initialize guides, videos, examples, ...
diff --git a/lib/jsduck/importer.rb b/lib/jsduck/importer.rb
index 50079cec646d8f3842c179ba6423172127c692ca..410dc0b468e0a0fc5f1b3833eefaa58a066fe8ce 100644
--- a/lib/jsduck/importer.rb
+++ b/lib/jsduck/importer.rb
@@ -9,20 +9,20 @@ module JsDuck
     module_function
 
     # Loads in exported docs and generates @since and @new tags based on that data.
-    def import(imports, relations)
+    def import(imports, relations, parallel)
       if imports.length > 0
-        generate_since_tags(read_all(imports), relations)
+        generate_since_tags(read_all(imports, parallel), relations)
       end
     end
 
     # Reads in data for all versions, returning array of
     # version/class-data pairs.  We don't use a hash to preserve the
     # order of versions (from oldest to newest).
-    def read_all(imports)
+    def read_all(imports, parallel)
       imports.map do |ver|
         {
           :version => ver[:version],
-          :classes => ver[:path] ? read(ver) : current_version,
+          :classes => ver[:path] ? read(ver, parallel) : current_version,
         }
       end
     end
@@ -32,14 +32,16 @@ module JsDuck
     end
 
     # Reads in data from all .json files in directory
-    def read(ver)
-      classes = {}
-      Dir[ver[:path] + "/*.json"].each do |filename|
+    def read(ver, parallel)
+      # Map list of files into pairs of (classname, members-hash)
+      pairs = parallel.map(Dir[ver[:path] + "/*.json"]) do |filename|
         JsDuck::Logger.instance.log("Importing #{ver[:version]}", filename)
         json = JsonDuck.read(filename)
-        classes[json["name"]] = members_id_index(json)
+        [json["name"],  members_id_index(json)]
       end
-      classes
+
+      # Turn key-value pairs array into hash
+      return Hash[ pairs ]
     end
 
     # creates index of all class members