# AI Automod
Automatically moderates messages the text messages in search of any harmful content.

This module scans all text messages in search of any harmful content. It can be fully customized to fit your requirements, but there's also 3 presets you can choose from.

### Global config
<style>
    .img-colm {
        display: flex;
        justify-content: center;
        align-items: center;
    }
    .img-colm img {
        max-width: 45vh;
        flex-shrink: 0;
        
    }

</style>
::: img-colm
-![]({DC694DA3-42EB-40DE-8D9F-37ADE28CD1D5}.png)

||| Light
Priories manual approval (enabled by setting a `Manual Approval channel`), allows swearing and light insults
||| Medium
A balance between both, swearing is allowed while heavy insults are flagged
||| Strict
Gives the power to flag most harmful content, in case you want clean environment ex. children around
||| 
:::

<br clear="both" />

### Manual-approval
In `Medium` and `Light` presets it is common to see messages pending for manual approval. It is meant for messages that are not 100% sure of its nature, so instead of automatically taking action, lets the moderator take action.
-![The manual approval report]({CEB3A9FD-28DE-434A-A9CC-886A3A7013A3}.png)
<br clear="both" />

[!badge variant="info" icon="person" text="Moderators"], can add a verdict to these reports by clicking on [!badge Add verdict]. Inside it there's few modules

Action                                                     | Description  
---                                                        | ---   
[!badge variant="danger" text="Punish"] |  Punishes the user. When clicked, a modal is open with settings
[!badge variant="dark" text="Accept without action"] | Accepts it without punishing the user (like I've read it)
[!badge variant="dark" text="Cancel actions"] | Cancels taking actions
[!badge variant="ghost" text="User info"] | Gets information about the user (WIP)
[!badge variant="base" text="False flag"] | Reports the message as a false flag