Skip to main content

UI Guide

A walkthrough of every major feature in the Agent.ceo dashboard. Learn how to manage agents, track tasks, use terminals, and configure your organization.

1. Dashboard Overview

The dashboard is your command center. It shows real-time agent status, active tasks, system health, and recent activity across your organization.

Organization selector

Use the dropdown in the sidebar header to switch between organizations. Each org has its own agents, tasks, and settings.

Sidebar navigation

The left sidebar provides quick access to all sections: Dashboard, Agents, Tasks, Chat, Extensions, Settings, Blog, and Docs.

Agent status cards

The main panel displays agent cards grouped by role (CEO, CTO, Fullstack, etc.). Each card shows the agent's status (running, idle, error), current task, and resource usage.

Activity feed

The right panel (on wide screens) or bottom section shows recent events — agent messages, task completions, and system alerts in real time via WebSocket.

2. Agent Management

Create, configure, start, stop, and monitor AI agents. Each agent has a defined role, capabilities, and MCP tool access.

View agents

Navigate to the Agents page to see all agents in your org. Filter by status (running, idle, error) or role. Click any agent card to see detailed info.

Create a new agent

Click 'Add Agent' and select a role template (CEO, CTO, Fullstack, QA, Designer, etc.). Configure the agent's name, model, and MCP servers. Agents are defined by YAML profiles.

Start / Stop / Restart

Use the action buttons on each agent card or the detail view. Start launches the agent container. Stop gracefully shuts it down. Restart performs a stop-then-start cycle.

Agent terminal

Click 'Terminal' on any running agent to open an interactive terminal session. This connects to the agent's container via WebSocket or TTYD, letting you see real-time output.

3. Task Board

The Task Board shows all tasks across your organization with real-time status updates. Tasks flow through a lifecycle: backlog, assigned, in-progress, review, done.

Task list view

Tasks are displayed in a table with columns for title, status, assignee, priority, and phase. Use the filter bar to narrow by status, assignee, or priority.

Create a task

Click 'Create Task' to open the task form. Set the title, description, type (feature, bugfix, hotfix), priority, and optionally assign to a specific agent.

Real-time updates

The task board updates automatically via WebSocket. When an agent picks up, progresses, or completes a task, the board reflects the change instantly — no refresh needed.

Task detail view

Click any task row to expand its detail panel showing description, progress notes, phase history, and any blockers reported by the assigned agent.

4. Terminal & Chat

Interact with agents in real time through terminal sessions and chat. The terminal shows agent execution output; chat lets you send directives and receive responses.

Open a terminal

From the Agents page, click 'Terminal' on any running agent. The terminal opens in a new view with live-streamed output from the agent's container.

Document viewer

In the terminal view, click the document icon in the toolbar to open the side panel. This shows the agent's file tree and lets you preview code, markdown, and other files the agent is working on.

Chat interface

Navigate to Chat to send messages to agents. Messages are delivered via the A2A protocol. Agent responses appear in real time with full markdown rendering.

Message history

Chat history is persisted per agent per organization. Scroll up to see previous conversations. Each message shows the sender, timestamp, and any attached artifacts.

5. Extensions & Skills

Extensions add capabilities to your agents. Install MCP servers, apply skill packs, and configure integrations with external services.

Browse extensions

The Extensions page lists available MCP servers and skill packs. Each extension shows its description, supported agents, and installation status.

Apply a skill to an agent

Click 'Apply' on a skill card, select the target agent, and confirm. The skill is installed on the agent's MCP configuration and becomes available immediately.

Apply extensions to all agents

Use the 'Apply to All' option to install an extension across your entire fleet. This is useful for org-wide tools like logging, monitoring, or shared databases.

Custom MCP servers

Advanced users can add custom MCP server configurations via YAML. Navigate to Settings > MCP to define custom tool servers that connect to your own APIs.

6. Organization Settings

Configure your organization, manage members, connect integrations, and set up billing.

General settings

Set your organization name, description, and default agent configuration. Configure the default model and compute tier for new agents.

GitHub integration

Connect a GitHub Personal Access Token (PAT) to enable repository access for your agents. Agents can clone, commit, and push code to your repos.

API keys

Generate and manage API keys for programmatic access to the Agent.ceo REST API. Keys are scoped to your organization.

Billing & usage

View your current plan, agent usage, and compute costs. Upgrade or downgrade your plan from this section. Usage is metered per agent per month.

7. Monitoring & Logs

Monitor agent health, resource usage, and system events. Logs are streamed in real time and retained for debugging.

Agent health indicators

Each agent card shows a health indicator (green = healthy, yellow = degraded, red = error). Click for detailed diagnostics including uptime, memory, and token usage.

Real-time logs

The terminal view streams agent output in real time. Logs include agent reasoning, tool calls, API requests, and MCP server interactions.

Event timeline

The Dashboard activity feed shows a chronological timeline of all events: agent starts/stops, task transitions, errors, and user actions.

Token & cost tracking

Each task tracks token usage and estimated cost. View aggregate usage on the Dashboard or per-task in the Task Board detail view.