Quick Start

This guide takes you from "fresh account" to "agent finished my first ticket" in about ten minutes.

The quick start is just the beginning. After these steps you will have a working baseline; but the real value — sharper instructions, more reliable agents, smoother human-agent collaboration — grows through continued use and refinement.

1. Create a Workspace

After signing in for the first time you land on the workspace onboarding screen.

1. Enter a workspace name (for example DBA Team)
2. Pick a slug — this becomes the URL segment (/dba-team/...). Slugs must be unique across Mopheus and cannot match reserved values (docs, api, admin, login, register, onboarding, invitations, workspaces, profile)
3. Click Create workspace

You are now the Owner of the new workspace and land on its dashboard.

2. Confirm a Runtime Is Ready

Agents need a runtime to execute. Open Runtimes in the sidebar and confirm at least one runtime shows status Idle.

Private deployment: If you are using the private deployment version, the runtime is already installed on the host and automatically logs in with the default account — no extra setup needed. To install additional runtimes on other machines, see Install Daemon.

If the list is empty or every runtime is Offline, no execution backend is currently available. You have two options:

• Wait for admin setup: If this is a team deployment, contact the administrator to confirm the runtime daemon is started
• Connect your own machine: Follow Install Daemon to install the Mopheus CLI on your own machine, start the daemon, and register it as an additional runtime

3. Invite an AI Colleague (Set Up an Agent)

Open Agents in the sidebar and click New agent.

1. Name: something memorable like DB Guardian
2. Runtime: pick one from the online runtimes list (this determines which provider CLI the agent uses to execute tasks)
3. Instructions: describe the agent's persona and rules. For example: "You are the DB Guardian, an experienced and cautious database expert. When given a ticket, first fully understand the requirement, then provide professional analysis, a feasible plan, and any necessary risk warnings. Write all conclusions as a ticket comment. For destructive operations (such as deleting data, changing core parameters, or restarting instances), you must explicitly flag the risk in the comment and wait for human confirmation before proceeding."
4. Click Create

4. Create a Ticket and Assign It to the Agent

From the sidebar, open Tickets and click New ticket.

1. Write a title — for example "Investigate yesterday's slow query root cause"
2. Add a description with context: the database name, time range, and directions already checked
3. Set the priority (Low, Normal, High, Urgent; Normal is the default, so leaving it unset equals Normal)
4. Set the status to Todo (an assigned ticket triggers agent execution unless its status is Backlog, Cancelled, or Archived — Todo, In Progress, In Review, Done, and Blocked all trigger)
5. Click Create

Once created, open the ticket, click the Assignee picker, switch to the Agents tab, and pick DB Guardian.

The server queues an agent task, a daemon claims it, and the agent starts working. You can watch the execution process and output live in the ticket's Runs tab.

---

At this point your AI colleague has finished your first piece of work. Everything below is optional expansion.

5. Invite Human Colleagues (Optional)

When you need humans to join the collaboration, from the sidebar open Settings → Members and click Invite.

1. Enter the invitee's email
2. Choose a role: Admin (can manage members and settings) or Member (everyday contributor)
3. Click Send invite

The invitee receives a link to /invitations/<token>. Once they accept, they appear in the members list.

6. Next Steps

• Skills — bundle reusable instructions into Skills and assign them to multiple agents
• Teams — group humans and agents into Teams and assign tickets to the whole team
• Jobs — automate recurring work with Jobs triggered by schedule, event, or webhook
• Inbox — subscribe to important tickets and get real-time notifications as state changes
• CLI — once you are familiar with the Mopheus CLI, you can create tickets, assign agents, and view run logs from the command line

---

Key tip: The quality of an agent's instructions and the quality of its skills are the single most important factors in execution quality — and the part of the workflow most worth iterating on. Do not expect a first-draft instruction to handle every case. Watch the first few runs carefully, then refine the wording, add constraints, and tighten boundary conditions based on actual output. That is how an agent becomes reliable.