You can get started creating an AI agent in just a few steps.
1Install and configure a Secure Agent.
A Secure Agent provides the runtime environment for your AI agent. You need a Secure Agent to enable your AI agent to connect to MCP servers, execute Python code in your AI agent, and deploy your AI agent.
2Create connections to the models, memory providers, and tools your AI agent will use.
3Define your AI agent's flow.
Open the agent flow canvas and drag an agent block into the workflow. Choose the LLM and tools your agent will use.
Optionally, add other agents and components to the workflow. Define variables to pass information between components.
4Deploy the agent flow to create your AI agent.
You can deploy your AI agent as a REST endpoint.
5Test, debug, and refine the workflow.
Use the test console in the agent flow canvas to interact with your agent and ensure that it produces the results you expect. You can also use a tool like Postman or cURL to test the agent. If you make changes to the agent, redeploy it so that the changes take effect.
After you deploy your agent, you can monitor its use. You can also organize your AI agents and tools into projects and folders and export and import them.