diff --git a/docs/workflow.md b/docs/workflow.md
index 55a8e66..42e8cdf 100644
--- a/docs/workflow.md
+++ b/docs/workflow.md
@@ -13,36 +13,35 @@ The crawler CLI will bulk-download feature data from pre-defined sources. The se
     }
 }%%
 sequenceDiagram
+    actor CLI
+    CLI->>Crawler: launch
     Crawler->>+NLDI-DB: Get Source Information
     Note left of NLDI-DB: SELECT * FROM nldi_data.crawler_source
     NLDI-DB-->>-Crawler: Sources table
 
-    Crawler->>+Feature_Source: Request Features
-    Note left of Feature_Source: HTTP GET ${crawler_source.source_uri}
-    Feature_Source-->>-Crawler: GeoJSON FeatureCollection
+    Crawler->>+FeatureSource: Request Features
+    Note left of FeatureSource: HTTP GET ${crawler_source.source_uri}
+    FeatureSource-->>-Crawler: GeoJSON FeatureCollection
     loop foreach feature in Collection
         Crawler-->>+Crawler: ORM
         Note right of Crawler: Parses and maps feature to SQL
         Crawler->>-NLDI-DB: Add to feature table
         Note left of NLDI-DB: INSERT INTO nldi_data.features
     end
-    Crawler->>+NLDI-DB: Relate Features
+    Crawler->>NLDI-DB: Relate Features
 
     %NLDI-DB-->>-Crawler: Success
 ```
+
 ## Annotations
 
-1) Connect to NLDI master database, requesting the list of configured feature sources.
-2) Returns a list of feature sources.  The crawler can either:
+1) Launch CLI tool
+2) Connect to NLDI master database, requesting the list of configured feature sources.
+3) Returns a list of feature sources.  The crawler can either:
    * list all sources and exit
    * Proceed to 'crawl' one of the sources in the table
-3) For the identified feature source, make a GET request via HTTP.  The URI is taken from the `crawler_sources` table.
-4) The feature source returns GeoJSON.  Among the returned data is a list of 'features'.
-
-For each feature:
-
-    5) Use the ORM to map the feature data to the schema reflected from the `features` table
-    6) Insert the new feature to the master NLDI database
-
-
-7) "Relate" features -- build the relationships matching features to their adjacent features in the NLDI topology.
\ No newline at end of file
+4) For the identified feature source, make a GET request via HTTP.  The URI is taken from the `crawler_sources` table.
+5) The feature source returns GeoJSON.  Among the returned data is a list of 'features'.
+6) **[Per-Feature]** Use the ORM to map the feature data to the schema reflected from the `features` table
+7) **[Per-Feature]** Insert the new feature to the master NLDI database
+8) "Relate" features -- build the relationships matching features to their adjacent features in the NLDI topology.
\ No newline at end of file