How to Add an /ask Slash Command That Calls AI in Discord

Mentioning a bot works, but slash commands feel cleaner: members type /ask, get an autocompleted prompt field, and the answer appears as a proper bot reply. This guide registers a global /ask command and uses a deferred reply so the model has time to respond without Discord timing out.

What you need

• A working Discord bot token and application ID
• Node 18+ and the discord.js library installed
• An AI API key
• The applications.commands scope checked when you invited the bot

Step 1: Register the command

Slash commands are registered with Discord's REST API once, not on every startup. Create a deploy script that pushes the command definition. Global commands can take up to an hour to appear, so use a guild-scoped registration while testing.

import "dotenv/config";
import { REST, Routes, SlashCommandBuilder } from "discord.js";

const commands = [
  new SlashCommandBuilder()
    .setName("ask")
    .setDescription("Ask the AI anything")
    .addStringOption((o) =>
      o.setName("prompt").setDescription("Your question").setRequired(true),
    )
    .toJSON(),
];

const rest = new REST().setToken(process.env.DISCORD_TOKEN);

await rest.put(
  Routes.applicationGuildCommands(
    process.env.APP_ID,
    process.env.GUILD_ID,
  ),
  { body: commands },
);
console.log("Registered /ask");

Step 2: Handle the interaction

In your main bot file, listen for InteractionCreate. Defer the reply immediately, run the AI call, then edit the original reply with the answer.

client.on(Events.InteractionCreate, async (i) => {
  if (!i.isChatInputCommand() || i.commandName !== "ask") return;

  await i.deferReply();
  const prompt = i.options.getString("prompt");

  const res = await ai.chat.completions.create({
    model: "gpt-5-mini",
    messages: [{ role: "user", content: prompt }],
  });

  await i.editReply(res.choices[0].message.content.slice(0, 1900));
});

Step 3: Test in your server

Restart the bot, type /ask in any channel, and fill in the prompt. You should see MyHelper is thinking... for a moment, then the answer replaces it.

Result

Members now have a native /ask command with an autocompleted prompt field. Because it uses a deferred reply, longer model responses never trigger the dreaded This interaction failed error.