Configurer un bot Discord de réponses automatiques par mot-clé

Mis à jour · 4 min de lecture

Utilisez le tableau de bord Cortex AI Pro pour créer un bot Discord qui répond par texte brut aux mots-clés. Une règle peut correspondre à un message exact, à un message qui commence par un déclencheur ou à un déclencheur présent dans le message. La casse est ignorée.

Commencer gratuitement

Commencez gratuitement. Pro coûte 9,99 USD par mois.

Avant de commencer

  1. Open Cortex AI settings for the correct server

    In the dashboard, open your bot, choose the Modules tab, then configure Cortex AI and select the intended server. Keep Enable auto-responses turned on. These dashboard rules are saved for the selected server.

    Select Add rule. A new dashboard rule starts with Exact message selected, while an omitted match type in saved settings defaults to Contains. Save Settings after making your changes so the bot receives the configuration.

  2. Add a trigger and choose the match type

    Enter the trigger, then choose Contains, Exact message, or Starts with. Contains looks for the trigger as a substring anywhere in the message; Exact message requires the whole message to equal the trigger; Starts with requires the message to begin with it. All three comparisons ignore letter case.

    Example — match a help request: use Contains with help to match “Need HELP please”; use Exact message with help to match only “help”; or use Starts with help to match “help me” but not “I need help”. These modes do not use word boundaries, so a Contains trigger can also match inside a longer word.

  3. Add reply variants and set a cooldown when needed

    Add one or more plain-text reply variants. A missing weight counts as 1; a positive whole-number weight makes that variant more likely to be selected. For example, weights 3 and 1 give the first reply three times the chance of the second over repeated matches. A weight does not guarantee a particular reply.

    Cooldown is measured in seconds. It defaults to 0, which means no cooldown, and its scope defaults to Per channel. With a channel cooldown, a matching rule stays silent for everyone in that channel until it expires; with Per user, that member stays silent across channels in the same server while other members can still receive a reply. A cooling match is handled silently rather than falling through to Cortex AI.

  4. Save and test with a member message

    Save Settings, then send a test message from a non-bot account in the selected server. Test the capitalization and message shape for the mode you chose. If the rule is cooling down, wait for it to expire or set Cooldown to 0 before trying again.

    If there is no reply, confirm Cortex AI and Enable auto-responses are enabled, you selected the correct server, and the bot can read the channel and receive message content. Dashboard keyword rules are evaluated before the older auto-responder records and Cortex AI chat.

Example — one help rule, three match choices

Match typeTriggerA matching member message
Containshelp“Need HELP please”
Exact messagehelp“help”
Starts withhelp“help me”

Questions fréquentes

Are Discord keyword auto-replies case-sensitive?

No. Contains, Exact message and Starts with all compare the trigger and message without letter case.

Can I make several replies for one keyword?

Yes. A dashboard rule accepts one through ten reply variants. Missing weights are equal; use positive whole-number weights to make one variant more likely.

What does Per user cooldown mean?

It suppresses that rule for the matching member across channels in the same server until the cooldown expires. Other members can still receive a reply.

Can I limit a keyword rule to a Discord role or channel?

No. Current dashboard rules offer channel or user cooldown scope, not per-trigger channel or role filters.

How many Discord keyword auto-reply rules can I save?

A server can save up to 50 dashboard rules. Each rule can use one through ten plain-text reply variants.

Prêt à configurer votre bot ?

Commencer gratuitement

Commencez gratuitement. Pro coûte 9,99 USD par mois.

Sources vérifiées

Les références de fichier et de ligne étayent les réglages et limites décrits.

  • server/templates/modules/cortex/replies.js:3-84
  • client/src/components/CortexReplySettings.tsx:37-103