diff --git a/README.md b/README.md
index 69bcf5b..7d3344e 100644
--- a/README.md
+++ b/README.md
@@ -9,6 +9,7 @@ Steps
 * run `python3 main.py`
   * required `--input <dir of input json>`
   * required `--output <dir of output csvs>`
+  * (optional) `--delimiter "|"`, change seperator/delimiter , default is |
   * (optional) to merge all json to single csv `--single`, default is one set of csv files for one json
   * (optional) # of records to check records for all headers `--metadata <number or records>` defaults to 1000
   * required `--join-column <column name from first level to use as merge column>`
diff --git a/main.py b/main.py
index bf258d6..d922013 100644
--- a/main.py
+++ b/main.py
@@ -10,6 +10,7 @@ import argparse
 parser = argparse.ArgumentParser()
 parser.add_argument("--input", help="folder containing input json(s)", required=True, type=pathlib.Path)
 parser.add_argument("--output", help="folder to place csv", required=True, type=pathlib.Path)
+parser.add_argument("--delimiter", help="delimiter for CSV (default is '|'", default="|")
 parser.add_argument("--single", action="store_true", help="merge all json files to single output csv")
 parser.add_argument("--metadata", type=int, help="how many records to parse for building metadata", default=1000)
 parser.add_argument("--join-column", help="join column from top-level to merge nested json", required=True)
@@ -65,7 +66,7 @@ class DBConn:
 
         for tbl in tbls:
             clients = pd.read_sql(f"SELECT * FROM \"{tbl}\"", self.con)
-            clients.to_csv(args.output / f"{prefix}{tbl}.csv", index=False)
+            clients.to_csv(args.output / f"{prefix}{tbl}.csv", index=False, sep=args.delimiter)
 
 
 dbConn = DBConn()