Getting Started
First Steps
Set up your CRHQ workspace — navigate the interface, create your first session, and start working with AI agents.
Once you're logged in, here's how to orient yourself and get productive quickly.
The Interface
CRHQ has four main areas:
- Sidebar (left) — Your sessions list, project folders, and navigation icons
- Main area (center) — The conversation view where you interact with agents
- Artifact panel (right, when active) — Visual outputs like code, documents, and HTML previews
- Bottom icons (sidebar footer) — Quick links to Tasks, Files, Account, and Settings
The sidebar has two viewing modes, toggled at the top:
- Sessions — A flat list of all your sessions, sorted by most recent
- Projects — Sessions grouped by project, with unassigned sessions at the bottom
Creating Your First Session
- Click the + New Session button at the top of the sidebar
- Choose an agent from the dropdown — each agent has different skills and specializations
- Optionally select a project to organize this session
- Give it a name (or let it auto-name based on the first message)
- Click Create
You're now in a conversation. Type a message and hit Enter (or click Send).
Choosing an Agent
CRHQ comes with pre-configured agents for common roles:
- Developer — Code, debugging, architecture, deployment
- Researcher — Web research, analysis, data gathering
- Content Writer — Marketing copy, blog posts, documentation
- General Assistant — Versatile, handles most tasks
Each agent has different skills assigned. You can see what an agent can do by going to Settings → Agents and clicking on one.
Tip: Not sure which agent to use? Start with the General Assistant — it can handle most tasks and delegate to specialized agents when needed.
Switching Models
By default, each agent uses its configured model (Haiku, Sonnet, or Opus). You can override this per-session using the model selector in the session header:
- Haiku — Fast responses, good for simple tasks
- Sonnet — Balanced speed and capability, the default choice
- Opus — Most capable, best for complex reasoning
Session Status
The session header shows the current status:
- Idle — Agent is ready for your input
- Running — Agent is actively working
- Error — Something went wrong (check the error message)
- Interrupted — Agent was stopped mid-task
Quick Tips
- Send follow-ups while the agent is working — messages queue and are delivered in order
- Stop an agent mid-task by clicking the Stop button in the header
- Search sessions using the search bar at the top of the sidebar
- Drag sessions to reorder them in the sidebar
- Click Info in the session header to see session details and auto-generated summary