> For the complete documentation index, see [llms.txt](https://imbrace.gitbook.io/imbrace-documentation/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://imbrace.gitbook.io/imbrace-documentation/ai-agent/start-a-new-agent.md).
# Start a New Agent
Create a New Agent from + button on the main AI Agent screen top right corner.
You can either create a new agent or import an agent. In order to create a new agent, click on the respective option, and choose from three agent types:
1. Orchestrator
**Purpose:** Coordinates multiple AI agents to work together toward a shared objective.
**Function:** Acts as a manager that assigns, sequences, and synchronizes tasks among different agents.
**Example:** An orchestrator that manages several AIs — one for customer service, one for billing, and one for logistics — ensuring they collaborate efficiently.
2. AI Agent
**Purpose:** Creates a smart assistant that can perform tasks and interact autonomously.
**Function:** Executes actions, responds to users, automates workflows, and handles assigned goals independently.
**Example:** A chatbot that schedules meetings, answers FAQs, or manages tasks on your behalf.
3. Guard Rail
**Purpose:** Defines rules, policies, and safety limits to keep AI behavior safe, focused, and compliant.
**Function:** Monitors and controls what AI agents can or cannot do; ensures ethical, legal, and secure operation.
**Example:** A system that enforces data privacy rules or restricts AI from generating harmful or unauthorized content.
