diff --git a/structure.md b/structure.md
new file mode 100644
index 0000000..84fe216
--- /dev/null
+++ b/structure.md
@@ -0,0 +1,127 @@
+# `main()`
+
+- Start timer
+- Parse command line
+- Build config object - `build_config()`
+- Setup logging
+- Open or create database
+- Select nitter instance to use
+- Get soup of whole page and timeline (list of soup of items) - `get_timeline()`
+- Iterate timeline to generate list of dicts with content of each tweet:
+    - Extract tweet ID
+    - Extract timestamp
+    - Skip if timestamp is not within acceptable range - `is_time_valid()`
+    - Skip if if it is a retweet and retweets are excluded
+    - Check database if tweet already exists and skip if it does
+    - Extract author name
+    - Extract twitter user name of author
+    - Extract full status page URL
+    - Add prefix if tweet is reply-to
+    - Add prefix if tweet is retweet
+    - Process media body `process_media_body()`
+    - Add link to quoted page ("card")
+    - Extract image(s) from card `process_card()`
+    - process video and image attachments `process_attachments()` 
+    - Add custom footer
+    - Add "Original tweet" footer
+    - Add optional timestamp to footer
+    - If no media, look for image in linked URL
+    - Get filename of downloaded video
+- Update user profile if necessary - `update_profile()`
+- Login to Mastodon instance - `login()`
+- Check toot character limit
+- Iterate list of tweets
+    - Check if toot cap not reached
+    - Upload video if applicable (previously downloaded)
+    - If no video and applicable, download and upload pic
+    - Find in database toot id of replied_to_tweet
+    - Post toot + insert in database
+- Clean up downloaded video files
+- Delete excess records in database
+
+# `build_config()`
+
+- Instanciate `global TOML` struct
+- populate TOML with default values
+- Load config file and (Over)write all valid keys with values read from file
+- If no config file, (Over)write all valid keys with values read from the command line
+- Verify that a minimum valid config is present
+ 
+# `get_timeline()`
+
+- Initiate requests session
+- Populate headers
+- Download nitter page of user
+- Make soup
+- Build a list with soup of each timeline item
+- Iterate list
+    - if individual tweet, add to final list
+    - if first tweet of thread, get the thread from tweet page - `_get_rest_of_thread()`
+
+# `_get_rest_of_thread()`
+
+- Download page
+- Make soup
+- Get all items in thread after main tweet
+- build list with references of previous tweet
+- Reverse timeline order
+
+# `is_time_valid()`
+
+- Compare timestamp to `tweet_delay` and `tweet_max_age` 
+
+# `process_media_body()`
+
+- Copy plain text
+- Convert links starting with @ and # to plain text
+- Remove redirection from links `deredir_url()`
+- Substitute source from links `substitute_source()`
+- Remove trackers from fragments `clean_url()`
+
+# `process_card()`
+
+- Get list of image URL in card tag
+
+# `process_attachments()`
+
+- Collect URLs of images
+- Download nitter video (converted animated GIF) and save it in output directory
+- Download twitter video by calling `youtube_dl` and save it in output directory
+
+# `update_profile()`
+
+- Extract banner and avatar picture addresses from soup
+- Get the banner and avatar picture addresses from database
+- If user record not found in db, create a new one
+- If they have changed
+    - Download banner and avatar pictures
+    - Login to Mastodon - `login()`
+    - Update credentials
+    - Record image URLs in database
+
+# `login()`
+
+- Create Mastodon application
+- Login with password if provided
+- Login with token
+
+# `deredir_url()`
+
+- Populate HTTP headers
+- Download the page
+
+# `substitute_source()`
+
+- Parse URL
+- Susbtitute domain values from config
+- Unparse URL
+
+# `clean_url()`
+
+- Parse URL
+- Remove UTM parameters from query and fragments
+- Unparse URL
+
+# `_remove_trackers_query(url_parsed.query)`
+
+# `_remove_trackers_fragment(url_parsed.fragment)`