> ## Documentation Index
> Fetch the complete documentation index at: https://docs.tuteliq.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Discord Integration

> AI-powered child safety monitoring for Discord servers

The Tuteliq Discord bot monitors text channels for grooming, bullying, and unsafe content in real time using the Tuteliq API. Add the hosted bot with one click or self-host on your own infrastructure.

<CardGroup cols={3}>
  <Card title="Add to Discord" icon="robot" href="https://discord.com/oauth2/authorize?client_id=1470719024320090154&scope=bot+applications.commands&permissions=2164260880">
    Invite the hosted bot to your server
  </Card>

  <Card title="Landing Page" icon="globe" href="https://tuteliq.ai/integrations/discord">
    Features overview and pricing
  </Card>

  <Card title="Self-Host" icon="github" href="https://github.com/Tuteliq/discord">
    Clone and deploy on your own infrastructure
  </Card>
</CardGroup>

## Installation

### Hosted bot (recommended)

Click the **Add to Discord** link above to invite the bot. No code required — configure everything with slash commands.

The bot needs the following permissions:

* Send Messages
* Embed Links
* Read Message History
* Manage Messages
* Moderate Members

Make sure **Message Content Intent** is enabled for the bot application.

### Self-host

```bash theme={"dark"}
git clone https://github.com/Tuteliq/discord.git
cd discord
npm install
cp .env.example .env
# Fill in your values (see Self-Hosting section below)
npm run deploy:commands
npm run build
npm start
```

## Quick start

<Steps>
  <Step title="Configure your API key">
    Run `/tuteliq setup` and enter your Tuteliq API key (starts with `tq_`). The key is encrypted at rest with AES-256.
  </Step>

  <Step title="Add channels to monitor">
    Run `/tuteliq monitor add #channel` for each text channel you want to protect.
  </Step>

  <Step title="Set the alert channel">
    Run `/tuteliq alerts channel #admin-alerts` to choose where safety alerts are posted.
  </Step>
</Steps>

## How it works

Messages in monitored channels flow through a multi-stage pipeline:

1. **Collect** — Each message is added to a per-channel sliding window buffer.
2. **Batch analyze** — The buffer is periodically flushed and sent to the Tuteliq API, which detects bullying, grooming, unsafe content, and emotional distress.
3. **Alert** — When a risk score exceeds your sensitivity threshold, a rich embed is posted to the alert channel with severity, risk score, message excerpt, and recommended action.
4. **Act** — Admins take action directly from the alert embed using interactive buttons.

## Slash commands

All commands live under the `/tuteliq` prefix and require **Administrator** or **Manage Server** permission.

| Command                             | Description                                             |
| ----------------------------------- | ------------------------------------------------------- |
| `/tuteliq setup`                    | Configure the bot with your Tuteliq API key             |
| `/tuteliq monitor add #channel`     | Start monitoring a text channel                         |
| `/tuteliq monitor remove #channel`  | Stop monitoring a channel                               |
| `/tuteliq monitor list`             | List all monitored channels                             |
| `/tuteliq alerts channel #channel`  | Set the channel for safety alerts                       |
| `/tuteliq alerts dm enable/disable` | Toggle admin DM notifications for high/critical threats |
| `/tuteliq alerts test`              | Send a test alert to verify setup                       |
| `/tuteliq sensitivity set <level>`  | Set detection sensitivity                               |
| `/tuteliq sensitivity view`         | View current sensitivity settings and thresholds        |
| `/tuteliq status`                   | View bot status, rate limits, and quota usage           |
| `/tuteliq link`                     | Generate a parent link code (expires in 15 minutes)     |
| `/tuteliq reset`                    | Remove all Tuteliq configuration from the server        |

## Alert actions

When a threat is detected, the alert embed includes interactive buttons:

| Action               | Effect                                                 |
| -------------------- | ------------------------------------------------------ |
| **Delete Message**   | Removes the flagged message from the channel           |
| **Timeout User**     | Temporarily mutes the user server-wide                 |
| **View Full Report** | Opens the detailed analysis with rationale and context |
| **Dismiss**          | Closes the alert without action                        |

Admins with DM notifications enabled also receive a direct message for high and critical severity threats.

## Sensitivity levels

| Level     | Threshold | Best for                                                         |
| --------- | --------- | ---------------------------------------------------------------- |
| `low`     | 70%+      | Minimal alerts, only clear violations                            |
| `medium`  | 50%+      | Balanced — good default for most servers                         |
| `high`    | 30%+      | Stricter monitoring, catches more subtle threats                 |
| `maximum` | 10%+      | Maximum protection — useful for testing or high-risk communities |

Set the level with `/tuteliq sensitivity set <level>`.

## Parent linking

The `/tuteliq link` command generates a one-time code that parents can enter in the Tuteliq mobile app to receive notifications when their child is involved in a flagged conversation. Link codes expire after 15 minutes.

## Self-hosting

### Requirements

* Node.js 20+
* Discord bot application with Message Content Intent enabled
* Tuteliq API key
* Firebase project with Firestore enabled
* Firebase service account key (JSON)

### Environment variables

```bash theme={"dark"}
# Discord Bot
DISCORD_TOKEN=your-bot-token
DISCORD_CLIENT_ID=your-client-id

# Tuteliq API
TUTELIQ_API_URL=https://api.tuteliq.ai

# Encryption (generate with: node -e "console.log(require('crypto').randomBytes(32).toString('hex'))")
ENCRYPTION_KEY=your-64-hex-char-encryption-key

# Firebase
GOOGLE_APPLICATION_CREDENTIALS=./service-account.json

# Server
PORT=8080
NODE_ENV=production
LOG_LEVEL=info
```

### Docker

```bash theme={"dark"}
docker compose up -d
```

### Cloud Run

```bash theme={"dark"}
gcloud builds submit --config=discord/cloudbuild.yaml .
```

### Health check

The bot exposes a health endpoint at `GET /health` on the configured port (default `8080`). It returns `200` when the bot is connected to Discord.

## Next steps

<CardGroup cols={2}>
  <Card title="API Reference" icon="code" href="/api-reference/introduction">
    Explore the full API specification.
  </Card>

  <Card title="Slack Integration" icon="slack" href="/integrations/slack">
    See the Slack integration for workspace chat protection.
  </Card>

  <Card title="Telegram Integration" icon="telegram" href="/integrations/telegram">
    See the Telegram integration for group chat protection.
  </Card>

  <Card title="Roblox Integration" icon="gamepad" href="/integrations/roblox">
    See the Roblox integration for in-game chat protection.
  </Card>
</CardGroup>