Set Up a Discord AI Chat Assistant With Cortex
Cortex AI can respond in selected Discord channels, with a server personality, conversation history controls and manager-only settings. It is a Pro module. Its generative replies need a compatible AI provider credential on the bot host, which is a hosting prerequisite rather than a Discord setting or a key entered by members.
Start freeStart free. Pro is $9.99/month.
Before you start
- A connected bot with Pro access and Cortex AI enabled.
- A compatible AI provider credential configured on the bot host by the hosting administrator; without one, generative replies are unavailable.
- A server manager with Manage Server permission to choose channels, personality and limits.
Enable Cortex and check the host prerequisite
Enable Cortex AI for the bot and use the normal regenerate and start workflow. In Discord, run /cortex to view the module settings. Before inviting members to rely on replies, confirm with the hosting administrator that a supported provider credential is configured for this bot host. This is not something to paste into the Discord Developer Portal or ask members to supply.
Cortex can use OpenAI or Anthropic credentials from its host environment. If neither is present, an AI response cannot be generated. Treat provider availability, network access and any host-level usage limits as operational prerequisites; this guide does not promise a particular model, cost or response time.
Select where Cortex can participate
Use /cortexchannel to add, remove or list AI channels. Keep access limited to channels where members expect the assistant to read and respond. The command is restricted to members with Manage Server, so ordinary members cannot add a new AI channel for the server.
Test by mentioning the bot in a selected channel after it is running. If it does not respond, check the module state, the selected channel, Discord permissions and the host’s provider configuration. Do not treat a missed response as evidence that a conversation was saved or answered elsewhere.
Set a personality and conservative limits
Use /cortexpersonality to set the assistant’s server personality and /cortexstyle to choose concise, balanced, detailed or creative responses. These are manager-only controls. Write a short, community-appropriate instruction and avoid putting secrets, private member data or instructions that should not leave your staff team into a prompt.
Use /cortexlimit to set a daily message limit per user; zero means unlimited in the command’s setting. Use /cortexbudget to set the module’s monthly budget in cents, and /cortexusage to inspect usage statistics. Host-level monthly limits can still apply, so a server setting is not a promise of unlimited AI calls.
Reset a channel’s conversation history when appropriate
Use /cortexreset in the current channel to clear that channel’s conversation history. Keep this control in your staff playbook for test conversations and for situations where continuing context is no longer appropriate.
Cortex settings and several optional AI features exist beyond chat, but do not enable an unfamiliar feature merely because the command exists. Review its setting, permissions and provider requirement before use. This guide covers channel chat setup; it does not promise that AI can replace moderators, guarantee factual answers or perform actions without the required permissions.
Frequently asked questions
Do members need their own OpenAI or Anthropic key?
No. The module reads compatible provider credentials from the bot host. A hosting administrator must configure them; members do not enter a key in Discord.
Can anyone enable Cortex in a channel?
No. Adding and removing AI channels, changing personality and setting usage controls require Manage Server permission.
Does Cortex have unlimited AI messages?
No. The bot host can impose a monthly call limit, and managers can set a daily per-user limit and a monthly budget. Provider availability also affects replies.
Can Cortex replace moderation or guarantee correct answers?
No. Treat responses as AI output that needs appropriate community oversight. Continue to use your existing moderation permissions and review processes.
