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

# Your First Agent

> Configure an agent and connect it to a messaging channel.

ClawHQ comes with several pre-built agents. This guide walks through activating one and connecting it to Discord.

## Activate an agent

1. Open the dashboard and go to **Team**
2. Click **Add Agent** or select an agent from the roster
3. The agent is now active — it will respond to messages in any connected channel

Each agent has a role, a set of tools, and a model. You can configure all three from the agent's settings page.

## Configure the agent

From **Team → \[Agent Name] → Settings**:

| Setting      | What it does                                                    |
| ------------ | --------------------------------------------------------------- |
| **Model**    | Which AI model the agent uses (Haiku, Sonnet, Opus)             |
| **Tools**    | What the agent can do (web search, code execution, file writes) |
| **Schedule** | Optional cron schedule for proactive tasks                      |
| **Memory**   | Whether the agent remembers past conversations                  |

### Choosing a model

| Model             | Best for                                   | Cost   |
| ----------------- | ------------------------------------------ | ------ |
| Claude Haiku 4.5  | Quick replies, summaries, triage           | \$     |
| Claude Sonnet 4.6 | Most tasks — balanced speed and capability | \$\$   |
| Claude Opus 4.6   | Complex reasoning, long-form writing       | \$\$\$ |

The [Model Router](/guides/model-router) can handle this automatically — routing each task to the right model based on what it is.

## Connect to Discord

See the full [Discord setup guide](/channels/discord), but the short version:

1. Go to **Settings → Channels → Discord**
2. Click **Connect Discord**
3. Follow the wizard to create a bot and paste the token
4. Invite the bot to your server using the generated invite link
5. Message the bot — it should respond

## Give the agent a task

Once connected, you can delegate tasks directly in Discord:

```
@Mike research the top 5 competitors to Notion and summarize their pricing
```

```
@Scout find recent news about OpenAI this week
```

Agents respond in the same channel. Long tasks update with progress messages.

## Built-in agents

| Agent      | What it does                                                                |
| ---------- | --------------------------------------------------------------------------- |
| **Hermes** | Conversational AI — persistent memory, long-term context, primary interface |
| **Mike**   | General-purpose CEO agent — handles anything                                |
| **Scout**  | Research and web intelligence                                               |
| **Codex**  | Code review, debugging, PR assistance                                       |
| **Scribe** | Writing, content, newsletters                                               |
| **Pixel**  | Design feedback, visual analysis                                            |
| **Fixer**  | Debugging, error investigation                                              |

<Note>
  **Hermes is different from the others.** It's the always-on conversational layer with persistent memory — it remembers context across every session. The agents above are specialized execution engines you trigger for specific tasks. See the [Hermes guide](/guides/hermes) for setup and usage.
</Note>

## Next steps

<CardGroup cols={2}>
  <Card title="Install a Pack" icon="box" href="/guides/packs">
    Add pre-configured agent bundles for your team
  </Card>

  <Card title="Model Router" icon="route" href="/guides/model-router">
    Automatically route tasks to the right model
  </Card>
</CardGroup>