Creating Flows

Flows are observable, traceable AI workflows. They wrap your AI logic to provide type safety, debugging support, and easy deployment as HTTP endpoints.

Defining a flow

// Simple flow with typed input/output
Flow<String, String, Void> greetFlow = genkit.defineFlow(
    "greeting",
    String.class,
    String.class,
    name -> "Hello, " + name + "!");

AI-powered flows

// Flow with AI generation and context access
Flow<String, String, Void> jokeFlow = genkit.defineFlow(
    "tellJoke",
    String.class,
    String.class,
    (ctx, topic) -> {
        ModelResponse response = genkit.generate(
            GenerateOptions.builder()
                .model("openai/gpt-4o-mini")
                .prompt("Tell me a short, funny joke about: " + topic)
                .build());
        return response.getText();
    });

Running flows

// Run a flow programmatically
String result = genkit.runFlow("greeting", "World");
// Returns: "Hello, World!"

Flows as HTTP endpoints

When using the Jetty or Spring Boot plugins, all defined flows are automatically exposed as HTTP endpoints:

# Call a flow via HTTP
curl -X POST http://localhost:8080/api/flows/tellJoke \
  -H "Content-Type: application/json" \
  -d '{"data": "pirates"}'

Why use flows?

• Type-safe inputs and outputs — Define clear schemas for your data
• Integrates with the Dev UI — Test and debug flows visually
• Easy deployment — Automatically exposed as HTTP endpoints
• Built-in tracing — Every flow execution is traced with OpenTelemetry
• Composable — Flows can call other flows

Next steps

Tool Calling Let models call your functions

Streaming Stream flow output in real-time

Dev UI & CLI Debug flows visually