Skip to main content
An agent is a graph of nodes on a canvas. Each node does one thing — take a text prompt, generate an image, edit a video — and you wire nodes together so the output of one feeds the input of the next. This walkthrough builds a minimal image agent: a text prompt that drives an image generator.

​
Build a text-to-image agent

1

Add a text prompt node

Open an agent in the studio and add a Text prompt node from the node toolbar. This node holds the text you want to generate from. Type your prompt directly into it, for example a paper crane on a warm beige background.
2

Add a generate image node

Add a Generate image node to the canvas. When a node has no connections it shows an idle state with a short description and a pulsing “Waiting for input” hint — that tells you it still needs something wired in.
3

Connect the nodes by their ports

Each node has typed ports: outputs on the right, inputs on the left. Drag from the text prompt’s text output to the generate image node’s prompt input. Ports are typed, so the canvas only lets you connect compatible ports — a text output connects to an input that accepts text. Once connected, an edge appears between the two nodes.
If a connection won’t take, the two ports have incompatible types. Check that you’re dragging from an output to an input that accepts the same kind of data.
4

Configure the image node

Select the generate image node to open its config panel. Pick a model and adjust settings like aspect ratio or number of variations. The estimated credit cost updates as you change the model and settings, so you can see what a run will cost before you trigger it.
5

Run the agent

Click Run. Pleyor saves your graph, then executes the nodes in order. The text prompt resolves first, then the generate image node starts. Running nodes show a generating animation while they work.
6

View the output

When the image node finishes, its output renders directly on the node. If you asked for multiple variations, each one appears as its own result. Click a node any time to inspect what it produced.

​
Next steps

Core concepts

Understand ports, edges, runs, and lifecycle states.

Running agents

Loop over many inputs and pause runs for human review.

Node reference

Every node type, its inputs, outputs, and settings.

Sharing your work

Share a read-only link or deploy the agent as an app.