diff --git a/README.md b/README.md
index ce3d6b4..b3c1ebb 100644
--- a/README.md
+++ b/README.md
@@ -4,28 +4,27 @@
 
 **This project is intended to run [this account](https://twitter.com/NijiHolo_EN_ID).**
 
+## Running modes
+The bot will be runnable in three modes:
+* Catch-up (`c`): intended to run only once, scan all accounts for cross-company tweets and post them. Terminate when done posting all.
+* Listen (`l`): post any new tweets that come up if it's cross-company
+    * may implement short-term catch-up function
+* Command-line: TBA
+
 ## Roadmap
-* Read past tweets of members from both companies
-* Track tweets in queue and history/log files
-* Create image of tweet, including up to three image attachments from that tweet
-* Combine image(s) with quote retweet
-* Don't tweet already-existing tweets (check our past quote RTs; might be saved in a file for quicker access)
-* Listen for live tweets as soon as they post
+* ~~Read past tweets of members from both companies~~
+* ~~Create screenshot of tweet~~
+* Track tweets in queue files
+* Post tweets along with screenshot of tweet
+    * Include screenshot of tweet being replied to?
+* Track posted tweets in a history file
+* Combine image with quote tweet
+* Don't tweet already-existing tweets (check our past quote tweets; might be saved in a file for quicker access)
 
 ## Notes
 * Tweets should only occur if involved parties are cross-company
 * Tweets should only occur if interaction involves [EN and EN] or [EN and (former) ID] parties
-    * cross-company ID interactions are regular enough
-* Text of our tweet should include involved names (w/o @) and the message, up to our tweet limit
-    * this, along with image, helps archive deleted tweets
-
-## API Calls
-**retrieving tweets from a user**
-
-https://oauth-playground.glitch.me/?id=usersIdTweets
-* `id`: user
-* `max_results=100`: 100 is the highest number of tweets we can retrieve
-* `pagination_token`: token that takes us to the next/prev page of tweets
-    * use `meta[next_token]`
-* `expansions=entities.mentions.username,in_reply_to_user_id`
-    * exposes mentions and replied users
\ No newline at end of file
+    * cross-company ID interactions are regular enough IMO
+* Our tweets should include:
+    * involved names (w/o @) and the tweet (as a quote tweet)
+    * screenshot of the tweet (+ maybe image of tweet being replied to when applicable)
\ No newline at end of file