diff --git a/batchdownload/launcher.py b/batchdownload/launcher.py
index 42d16e3..0682b2d 100644
--- a/batchdownload/launcher.py
+++ b/batchdownload/launcher.py
@@ -77,9 +77,9 @@ for wiki in wikis:
     # typically they don't provide any crawl-delay value in their robots.txt).
     if started and wikidir: #then resume
         print 'Resuming download, using directory', wikidir
-        subprocess.call('python dumpgenerator.py --api=%s --xml --images --resume --path=%s' % (wiki, wikidir), shell=True)
+        subprocess.call('./dumpgenerator.py --api=%s --xml --images --resume --path=%s' % (wiki, wikidir), shell=True)
     else: #download from scratch
-        subprocess.call('python dumpgenerator.py --api=%s --xml --images' % wiki, shell=True)
+        subprocess.call('./dumpgenerator.py --api=%s --xml --images' % wiki, shell=True)
         started = True
         #save wikidir now
         for dirname, dirnames, filenames in os.walk('.'):