diff --git a/native_client/coqui-stt.h b/native_client/coqui-stt.h
index a08e3ab4..6140bf6d 100644
--- a/native_client/coqui-stt.h
+++ b/native_client/coqui-stt.h
@@ -329,7 +329,10 @@ STT_EXPORT
 char* STT_IntermediateDecodeExpensive(StreamingState* aSctx);
 
 /**
- * @brief Compute the intermediate decoding of an ongoing streaming inference,
+ * @brief Compute the intermediate decoding of an ongoing streaming inference, flushing
+          buffers first. This ensures that all audio that has been streamed so far is
+          included in the result, but is more expensive than STT_IntermediateDecode()
+          because buffers are processed through the acoustic model.
  *
  * @param aSctx A streaming state pointer returned by {@link STT_CreateStream()}.
  * @param aNumResults The number of candidate transcripts to return.
diff --git a/native_client/python/__init__.py b/native_client/python/__init__.py
index ee5685bd..751a3d0d 100644
--- a/native_client/python/__init__.py
+++ b/native_client/python/__init__.py
@@ -283,8 +283,10 @@ class Stream(object):
 
     def intermediateDecodeExpensive(self):
         """
-        Compute the intermediate decoding of an ongoing streaming inference, flushing buffers.
-        This ensures that all data that has been streamed so far are included in the result.
+        Compute the intermediate decoding of an ongoing streaming inference, flushing
+        buffers first. This ensures that all audio that has been streamed so far is
+        included in the result, but is more expensive than intermediateDecode() because
+        buffers are processed through the acoustic model.
 
         :return: The STT intermediate result.
         :type: str