Home / Docs / Streamer docs / Console walkthrough
Console walkthrough
The console has six steps, shown as tabs along the top. Here is what each one does.
Streamer docs. You run a Twitch channel and want your mods ready for a raid. Not what you came for? Find the right docs.
You can move between steps at any time. The mockups below use sample data from the demo channel; yours shows your own filters, chat and grades. You can also click through the console preview.
1. Read filters
What your channel blocks today, read from Twitch. ModScrim tests these and never changes your AutoMod settings.
- Public blocked terms. Twitch doesn’t share private blocked terms with apps, so they aren’t shown or tested. Refresh from Twitch reads them again.
- Chat log. The number of messages, when collecting started, and that they are kept for 7 days. Recent lines are blurred until you hover or focus them. Delete chat log and Pause logging are here.
- AutoMod levels. Read only. Change these in your Twitch moderation settings.
2. Configure
The join code already works while you change settings, so mods can join now.
- Tactics: Fake giveaway links, Follower and viewer spam, Art and promo bots, and Hate raid (placeholder slurs), which is off by default and shows a content warning.
- Numbers: attack lines (6 to 30, at most 40 lines in total), banter lines (4 to 20), seconds between lines (0.8 to 3.0) and the slip window (5 to 15 seconds).
- Kai (sim), a simulated co-mod, always labelled and scored separately.
- Invite your mods: the join code and the mod link. OBS overlay: the overlay URL, Rotate key, Preview the overlay and Hide mod names on the overlay.
3. Run the drill
The same practice chat your mods see. You can act too, or just watch the scoreboard. Press Start drill. Anything nobody removes within the slip window slips through.
4. Twitch grades
Every attack line from the drill is checked against your current filters with Twitch’s Check AutoMod Status. You see one plain number and a line-by-line table with the tactic and trick each line used. See reading the grades.
5. Fix the rules
Terms that catch what got past, checked against your last 7 days of chat and the banter pool. Any term that would match a regular is rejected.
- Proposed: up to six terms, each with its real-chat matches (must be zero) and an example line it catches. Untick any you don’t want.
- Add your own term: 2 to 500 characters; a
*may appear only at the start or end of a word. - Rejected: terms that would ban a regular, with the line they matched.
- Needs a bot rule: tricks blocked terms can’t express, such as spaced letters. Use a moderation bot for those.
- Catches 2
giftsnow*0 real-chat matches - Rejected: matches a regular’s “dm me the build later”
dm meWould ban a regular
6. Fresh wave and push
30 new attacks from a new seed, built with tricks the rule finder never saw. You see how many your filters block before and after, what still gets through, and that the new rules matched zero real messages. Then Approve and push sends the terms you ticked to Twitch. Undo this push removes them. See approve, push and undo.