diff --git a/README.md b/README.md
index 4ee9bb9..c2c07ff 100644
--- a/README.md
+++ b/README.md
@@ -62,14 +62,15 @@ To see all available options type <code>typist --help</code> or just
 ### Tips
 
 Doing the same test over and over again isn't good practice. I have provided
-a text file and a small python script to randomize the most used words in the
-English language. You can use this script on any text file, though it will
-remove punctuation marks and reformats the text.
+a text file and a small python script and a shell script to randomize the most
+used words in the English language. You can use this script on any text file,
+though it will remove punctuation marks and reformats the text.
 
 ```
-python3 utils/randomize.py most-used-words.txt
+python3 utils/randomize.py most_used_words.txt
+# or
+utils/randomize.sh most_used_words.txt
 ```
-
 ---
 
 ### [Issues / Bugs](https://github.com/ny64/typist/issues)
diff --git a/utils/randomize.sh b/utils/randomize.sh
new file mode 100755
index 0000000..8373237
--- /dev/null
+++ b/utils/randomize.sh
@@ -0,0 +1,50 @@
+#!/bin/sh
+
+# use a POSIX shell shuf alternative if none on the system
+command -v shuf > /dev/null 2> /dev/null || shuf() {
+    awk '
+        BEGIN {
+            srand();
+            OFMT = "%.17f"
+        }
+        {
+            print rand(), $0
+        }
+    ' "$@" | 
+    sort -k1,1n | 
+    cut -d ' ' -f2- ;}
+
+# first argument is the filename
+filename="$1"
+[ ! -f "$filename" ] && printf "%s\n" "You must provide a text file as argument." >&2 && exit 1
+
+words=""
+# read file line by line
+while IFS= read -r line; do
+    # split the line into words and set them as positional parameters
+    # shellcheck disable=SC2086
+    set -- $line
+    for val; do
+        # clean up each word: remove leading and trailing punctuation and quotes
+        val=$(printf "%s" "$val" | sed -e 's/^[({'"'"'"]*//' -e 's/[.,!?;:)]}'"'"'"]*$//')
+        words="$words
+$val"   # one word per line in $words var
+    done
+done < "$filename"
+
+line_length=0
+{
+    # loop over each shuffled words
+    printf "%s\n" "$words" | shuf | while IFS= read -r word; do
+        # update line length
+        line_length=$((line_length + ${#word} + 1))
+        # if line length exceeds 80 characters, make a new line
+        if [ "$line_length" -gt 80 ]; then
+            echo
+            # update line length
+            line_length=$(( ${#word} + 1 ))
+        fi
+        printf "%s " "$word"
+    done
+echo
+} > "$filename"