diff --git a/q2_fragment_insertion/_insertion.py b/q2_fragment_insertion/_insertion.py
index 710164b..64347dc 100644
--- a/q2_fragment_insertion/_insertion.py
+++ b/q2_fragment_insertion/_insertion.py
@@ -19,6 +19,7 @@
                                    DNAFASTAFormat,
                                    DNAIterator)
 from q2_types.tree import NewickFormat
+from qiime2.plugin import get_available_cores
 
 from q2_fragment_insertion._format import PlacementsFormat, SeppReferenceDirFmt
 
@@ -61,6 +62,9 @@ def sepp(representative_sequences: DNASequencesDirectoryFormat,
          debug: bool = False,
          ) -> (NewickFormat, PlacementsFormat):
 
+    if threads == 0:
+        threads = get_available_cores()
+
     placements = 'q2-fragment-insertion_placement.json'
     tree = 'q2-fragment-insertion_placement.tog.relabelled.tre'
 
diff --git a/q2_fragment_insertion/plugin_setup.py b/q2_fragment_insertion/plugin_setup.py
index f88f694..21a04ab 100644
--- a/q2_fragment_insertion/plugin_setup.py
+++ b/q2_fragment_insertion/plugin_setup.py
@@ -39,7 +39,7 @@
         'reference_database': SeppReferenceDatabase,
     },
     parameters={
-        'threads': qiime2.plugin.Int,
+        'threads': qiime2.plugin.Threads,
         'alignment_subset_size': qiime2.plugin.Int,
         'placement_subset_size': qiime2.plugin.Int,
         'debug': qiime2.plugin.Bool,
@@ -55,7 +55,8 @@
                               'representative sequences into.',
     },
     parameter_descriptions={
-        'threads': 'The number of threads to use.',
+        'threads': 'The number of threads to use. Pass 0 to use one per '
+                   'available core.',
         'alignment_subset_size': 'Each placement subset is further broken '
                                  'into subsets of at most these many '
                                  'sequences and a separate HMM is trained on '