diff --git a/streamrip/config.py b/streamrip/config.py
index c3c2c64..247c5fe 100644
--- a/streamrip/config.py
+++ b/streamrip/config.py
@@ -3,6 +3,7 @@ import copy
 import logging
 import os
 import re
+from functools import cache
 from pprint import pformat, pprint
 
 from ruamel.yaml import YAML
@@ -15,6 +16,7 @@ from .constants import (
     TRACK_FORMAT,
 )
 from .exceptions import InvalidSourceError
+from .utils import safe_get
 
 yaml = YAML()
 
@@ -149,6 +151,9 @@ class Config:
             logger.debug("Config saved: %s", self._path)
             yaml.dump(info, cfg)
 
+        docs = ConfigDocumentation()
+        docs.dump(self._path)
+
     @property
     def tidal_creds(self):
         """Return a TidalClient compatible dict of credentials."""
@@ -194,20 +199,17 @@ class Config:
         return f"Config({pformat(self.session)})"
 
 
-class ConfigDocumentationHelper:
-    """A helper class that writes documentation for the config file.
+class ConfigDocumentation:
+    """Documentation is stored in this docstring.
     qobuz:
         quality: 1: 320kbps MP3, 2: 16/44.1, 3: 24/<=96, 4: 24/>=96
         app_id: Do not change
         secrets: Do not change
     tidal:
         quality: 0, 1, 2, or 3
-        user_id: Do not change
-        country_code: Do not change
-        access_token: Do not change
-        refresh_token: Do not change
-        token_expiry: Do not change
-    deezer: Does not require login
+        user_id: Do not change any of the fields below
+        token_expiry: Tokens last 1 week after refresh. This is the Unix timestamp of the expiration time.
+    deezer: Deezer doesn't require login
         quality: 0, 1, or 2
     soundcloud:
         quality: Only 0 is available
@@ -233,32 +235,75 @@ class ConfigDocumentationHelper:
         track: Available keys: "tracknumber", "artist", "albumartist", "composer", and "title"
     lastfm: Last.fm playlists are downloaded by searching for the titles of the tracks
         source: The source on which to search for the tracks.
-    concurrent_downoads: Download (and convert) tracks all at once, instead of sequentially. If you are converting the tracks, and/or have fast internet, this will substantially improve processing speed.
+    concurrent_downloads: Download (and convert) tracks all at once, instead of sequentially. If you are converting the tracks, and/or have fast internet, this will substantially improve processing speed.
     """
 
-    comments = _set_to_none(copy.deepcopy(Config.defaults))
-
     def __init__(self):
         self.docs = []
         doctext = self.__doc__
+        # get indent level, key, and documentation
         keyval = re.compile(r"( *)([\w_]+):\s*(.*)")
         lines = (line[4:] for line in doctext.split("\n")[1:-1])
 
         for line in lines:
             info = list(keyval.match(line).groups())
             if len(info) == 3:
-                info[0] = len(info[0]) // 4
-            else:
+                info[0] = len(info[0]) // 4  # here use standard 4 spaces/tab
+            else:  # line doesn't start with spaces
                 info.insert(0, 0)
 
             self.docs.append(info)
-        pprint(self.docs)
 
-    def write(self):
-        pass
+    def dump(self, path: str):
+        """Write comments to an uncommented YAML file.
 
-    def __setitem__(self, key, val):
-        self.comments[key] = val
+        :param path:
+        :type path: str
+        """
+        is_comment = re.compile(r"^\s*#.*")
+        with open(path) as f:
+            # includes newline at the end
+            lines = f.readlines()
 
-    def __getitem__(self, key):
-        return self.comments[key]
+        with open(path, "w") as f:
+            while lines != []:
+                line = lines.pop(0)
+                found = False
+                to_remove = None
+                for level, key, doc in self.docs:
+                    # using 1 indent = 2 spaces like ruamel.yaml
+                    spaces = level * "  "
+                    comment = f"{spaces}# {doc}"
+
+                    if is_comment.match(line):
+                        # update comment
+                        found = True
+                        break
+
+                    re_obj = self._get_key_regex(spaces, key)
+                    match = re_obj.match(line)
+                    if match is not None:  # line contains the key
+                        if doc != "":
+                            f.write(f"{comment}\n{line}")
+                            found = True
+                        to_remove = [level, key, doc]
+                        break
+
+                if not found:  # field with no comment
+                    f.write(line)
+
+                if to_remove is not None:
+                    # key, doc pairs are unique
+                    self.docs.remove(to_remove)
+
+    @cache
+    def _get_key_regex(self, spaces, key):
+        regex = rf"{spaces}{key}:(?:$|\s+?(.+))"
+        return re.compile(regex)
+
+    def strip_comments(self, path: str):
+        with open(path, "r") as f:
+            lines = [line for line in f.readlines() if not line.strip().startswith("#")]
+
+        with open(path, "w") as f:
+            f.write("".join(lines))