Help Center Welcome, Guest Login

How to Use AutoMod

Last Updated: Mar 25, 2017 10:42PM PDT

How to Use AutoMod

This article explains how to use AutoMod to assist in chat moderation, as well as other tips and answers to frequently asked questions.

Notice an error or have feedback for this article? Let us know!
To improve moderation experience and broadcaster support even further, you can now shield your chat with AutoMod. AutoMod uses machine learning and natural language processing algorithms to hold risky messages from chat so they can be reviewed by a channel moderator before appearing to other viewers in the chat. Moderators can approve or deny any message caught by AutoMod. With AutoMod, broadcasters can establish their own reliable baseline for acceptable language and around-the-clock chat moderation, a way to create a more positive experience for themselves and their communities.

Index (click to jump to):
How AutoMod Works
What Each Level Means
How to Enable AutoMod
Frequently Asked Questions


AutoMod is a unique moderation tool that does more than filter inappropriate chat. When a user sends a message that AutoMod flags as potentially inappropriate, the message is held in a queue awaiting moderator approval. AutoMod also enables broadcasters to adjust the degree of filtering in the event they are more, or less, conservative about the type of dialogue they want to see in their chat. Beyond identifying inappropriate words and phrases, AutoMod can detect misspellings of prohibited words, potentially inappropriate strings of emotes and other characters or symbols that others could use to evade filtering.

Broadcasters can configure AutoMod by selecting one of four levels. These levels affect how aggressively AutoMod holds back messages under four categories: identity, sexual language, aggressive speech, and profanity. These settings can only be changed by the broadcaster, while the dictionary each setting contains is managed by Twitch.

AutoMod does not timeout, ban, or mute users from any channel; it merely withholds comments that fall within the broadcaster's chosen moderation settings to be either approved or denied by the broadcaster's moderators.

Users will receive a notification that their message is being held for review. All users will receive instant feedback that their message is being held for review, as well as a notification if the message was approved or denied. Of course, if approved, the message will instantly appear in chat as the next message.


Broadcasters and their moderators will also still see all the same timeout and ban chatbox options on those messages, so even if the message was denied, there will still be the ability to timeout or ban the user.

You can choose one of four levels for your AutoMod setting. Each level catches all words from the previous level, plus more. The higher the level, the more AutoMod will catch. You can see how strictly AutoMod will filter based on how many shields are displayed in each language category, from one (a little) to four (a lot).

AutoMod language categories:
  • Identity language - Words referring to race, religion, gender, orientation, disability, or similar. Hate speech falls under this category.
  • Sexually explicit language - Words or phrases referring to sexual acts, sexual content, and body parts.
  • Aggressive language - Hostility towards other people, often associated with bullying.
  • Profanity - Expletives, curse words, and vulgarity. This filter especially helps those who wish to keep their community family-friendly.
What AutoMod catches at each level:
  • Level 1: Only remove hate speech.
  • Level 2: Also remove sexually explicit language and abusive language.
  • Level 3: Remove even more hate speech and sex words.
  • Level 4: All of the above, plus profanity and mild trash talk.

We recommend you discuss with or inform your moderator team what settings you have used and perhaps which words or phrases you care most about filtering, in order to ensure they adhere to your preferences when they are prompted by AutoMod to review chat messages for approval or denial.

To configure AutoMod:
  1. Go to your Channel Settings page and scroll down to the Chat Options section.
  2. Click on the “Start setting up AutoMod” button.
  3. You will be defaulted to Level 1, the lowest moderation setting. Drag the slider bar to choose the setting most appropriate for you, from Level 1 to Level 4. Review the other settings in the Chat Options section of your Channel Settings to ensure everything is set up or updated to work best for your community.
  4. Scroll to the bottom of the page and click “Save.”
While there is no time limit on how long a comment can be pending without being approved or denied, the message will disappear from chat as the same rate as they clear overall, which is  a maximum of 200 messages in chat at any given time. This means moderators will need to review pending messages before they scroll out of chat.
  • Q: What languages is AutoMod available in?
    • A: The full AutoMod tool is currently available in English.
  • Q: Are there plans to expand the languages available for AutoMod?
    • A: Twitch plans to expand the list of available languages as development continues. Announcements for further available languages will be made on this article and on Twitch's Twitter account.
  • Q: How will AutoMod stay up to date as new meanings and slang words grow on the internet so frequently?
    • A: Twitch will be regularly reviewing how moderators use AutoMod to maintain its effectiveness on a site-wide and per channel basis. The feedback from your moderators approving and denying messages, as well as removing other messages not caught by AutoMod, will help to grow our understanding of how language used on Twitch evolves. Our goal is for a broadcaster's chosen AutoMod setting to reflect how they want to run their chat.
  • Q: What other tools and options are there to ensure a quality chat experience for viewers?
  • Q: What if I want to customize my AutoMod settings even further, like with a whitelist or blacklist for words the settings I have chosen do not include?
    • A: Whitelisting and blacklisting abilities are coming soon!