Guides

Company Brain — Complete User Guide

Main Page / Overview

Path: Frontline → Agents → Company Brain

What it does

Company Brain is where you create and manage AI assistants for your organization. It includes two main assistant types: OpenClaw Assistant, a multi-platform AI with memory and community skills, and Hermes Agent, a self-improving AI that learns workflows and creates skills over time.

Main elements

• Left sidebar: Agents module with Company Brain selected.
• Sub-menu: OpenClaw Assistant, Hermes Agent, Social / Postiz, OpenDesign Studio, Supermemory Engine, and Paperclip Orchestrator.
• Main area: Lists existing assistants or shows an empty state.
• Add button: Creates a new assistant.

Creating an OpenClaw Assistant

Click + Add OpenClaw Assistant to start. In the first step, fill in:

• Setup Type — Choose to create a new server or link an existing assistant server from another SaaS.
• Identifier Name — A unique name for the assistant.
• Description — What the assistant will be used for.

Click Continue to proceed.

Choosing an AI Provider

In the next step, select the AI provider and model. Available providers include Kimi For Coding, Moonshot Kimi, OpenAI, Anthropic Claude, Google Gemini, OpenRouter, DeepSeek, xAI Grok, and GLM (Z.AI).

After selecting a provider, enter the required API key and choose a model from the dropdown. Click Connect erxes AI Assistant to continue.

Provisioning the Assistant

The system creates a private runtime for the assistant. This process usually takes 2–4 minutes and includes:

• Preparing deployment
• Creating a private runtime
• Starting the assistant
• Installing tools and workspace
• Verifying runtime health

When the runtime is ready, you can click Open assistant to start using it.

Assistant Card

After provisioning, the assistant appears as a card showing:

• Status — Ready.
• Edit assistant — Change assistant settings.
• Connect Discord — Link the assistant to a Discord server.
• Name and description — The identifier and purpose.
• Provider — The connected AI provider.
• Invited members — Team members with access.

Click Open AI Assistant to launch the chat interface.

Connecting Discord

Click Connect Discord to make the assistant available in a Discord server. The panel shows:

• Runtime status — Ready or another status.
• Server and Server ID — The managed assistant server.
• Discord server — Select the server to install the assistant into.
• Discord channel — Choose the channel where the assistant will respond.

Click Connect channel to finish.

Using the Assistant in Discord

Once connected, the assistant appears in the selected Discord channel. Users can mention or message the assistant, and it will reply in the channel.

Opening the AI Assistant

Back on the assistant card, click Open AI Assistant to use the built-in chat interface.

AI Assistant Chat Interface

The chat interface is where you interact with the assistant. It shows:

• Top bar: Refresh, members, files, tools, share, and delete options, plus the active model dropdown.
• Main chat area: A welcome message and suggested prompts such as:
  • What can you do?
  • Summarize my recent sessions
  • Help me configure a channel
  • Check system health
• Message input: Type a message or use the microphone button for voice input.

Assistant Response and Usage Stats

When you send a message, the assistant replies. A side panel shows:

• Context window — Current usage compared to the limit.
• Latest run tokens — Input, output, and estimated cost.
• Cost by type — Breakdown of input, output, cache read, and cache write costs.
• Provider and model — The active provider and model name.

• 

Model Settings

Click the model selector at the bottom of the chat to adjust settings. The panel lets you:

• Change the provider.
• Adjust reasoning level on a scale from Faster to Smarter.
• Set the speed — Default, Fast, Standard, or Auto.
• Choose whether to use the default model.
• Save or discard changes.