Build an agent

This section covers how to build, deploy, and run agentic AI applications on Flyte. You’ll learn how to implement common agent patterns like ReAct and Plan-and-Execute and deploy agents as hosted services.

Quickstart

Here’s how Flyte maps to the agentic world:

• TaskEnvironment: The sandboxed execution environment for your agent steps. It configures the container image, hardware resources (CPU, GPU), and secrets (API keys). Think of it as defining “where this code runs.”
• @env.task: Turns any Python function into a remotely-executed step. Each task runs in its own container with the resources you specified. This is the equivalent of a node in LangGraph or n8n.
• Tasks calling tasks: A task can await other tasks, and each called task gets its own container automatically. No separate workflow decorator needed. The calling task IS your workflow, this is how you build multi-step agentic pipelines.
• @flyte.trace: Marks helper functions inside a task for fine-grained observability and caching. Each traced call appears as a span in the Flyte dashboard, with its inputs and outputs captured and checkpointed. Use this on your LLM calls, tool executions, and routing decisions to get full visibility into every turn of the agent loop.

Next steps

• Deploy an agent as a service: Host a FastAPI app, webhook pattern, model serving
• Building agents on Flyte: ReAct pattern, Plan-and-Execute with fan-out, LangGraph integration, and more patterns