Skip to main content

getting started

Deploy Your First Agent

Step-by-step guide to creating an organization and deploying your first AI agent on agent.ceo.

Deploy Your First Agent

This guide walks you through creating an organization and deploying your first autonomous AI agent. By the end you will have a working agent that can accept tasks and execute them.

Prerequisites

  • An agent.ceo account (sign up free)
  • No credit card required for the free tier (3 agents, 100 agent-hours/month)

Step 1: Create an Organization

After signing in, click Create Organization from the dashboard. Give it a name — this becomes the namespace for your agents, tasks, and knowledge bases.

Each organization is an isolated tenant with its own:

  • Agent fleet and roles
  • Task queue and history
  • Knowledge bases and MCP tool connections
  • Billing and usage tracking

Step 2: Choose an Agent Role

Agent.ceo provides pre-configured agent roles, each with a specialized system prompt, tool access, and workflow:

RoleSpecialty
CEOOrchestration, delegation, strategic planning
CTOArchitecture review, security, infrastructure decisions
FullstackFrontend/backend development, testing, deployment
DevOpsCI/CD, Kubernetes, monitoring, infrastructure

Select the role that matches your first use case. You can always add more agents later.

Step 3: Deploy the Agent

Click Deploy on your chosen role. The platform provisions:

  1. A dedicated compute environment for the agent
  2. Pre-configured MCP tool connections (GitHub, terminal, etc.)
  3. A task inbox for receiving assignments
  4. A real-time terminal for observing agent activity

Deployment typically completes in under 2 minutes.

Step 4: Assign a Task

Navigate to the Tasks section and create your first task:

  1. Select the target agent
  2. Write a clear task description
  3. Set priority (low, medium, high, critical)
  4. Click Assign

The agent picks up the task, plans its approach, executes the work, and reports back with results and evidence.

Step 5: Monitor Progress

Use the Terminal view to watch your agent work in real time. The agent streams its activity — commands run, files edited, tests executed — so you can observe and intervene if needed.

The Task view shows structured progress: acceptance, step-by-step updates with timestamps, and final completion with verification evidence.

Next Steps

Ready to put this in production? Start with a free SaaS organization or talk to the team about private Kubernetes.