Generating Content

Genkit provides a unified interface for working with generative AI models from any supported provider. Configure a model plugin once, then call any model through the same API.

The generate() method

The primary interface for interacting with AI models is generate():

ModelResponse response = genkit.generate(
    GenerateOptions.builder()
        .model("openai/gpt-4o-mini")
        .prompt("Invent a menu item for a pirate-themed restaurant.")
        .build());

System.out.println(response.getText());

Model parameters

Control how the model generates content with configuration options:

ModelResponse response = genkit.generate(
    GenerateOptions.builder()
        .model("openai/gpt-4o-mini")
        .prompt("Tell me a creative story.")
        .config(GenerationConfig.builder()
            .temperature(0.9)       // 0.0-2.0, higher = more creative
            .maxOutputTokens(512)   // Limit output length
            .topP(0.95)             // Nucleus sampling
            .topK(40)               // Top-K sampling
            .build())
        .build());

Parameter reference

Parameter	Description
temperature	Controls randomness. Low values (0.0-1.0) = more deterministic, high values (>1.0) = more creative
maxOutputTokens	Maximum number of tokens to generate
topP	Nucleus sampling — cumulative probability threshold (0.0-1.0)
topK	Limits token selection to the top K most likely tokens
stopSequences	Character sequences that signal the end of generation

Switching models

Changing models is as simple as changing the model string:

// Use OpenAI
genkit.generate(GenerateOptions.builder()
    .model("openai/gpt-4o")
    .prompt("Hello!").build());

// Use Gemini
genkit.generate(GenerateOptions.builder()
    .model("googleai/gemini-2.0-flash")
    .prompt("Hello!").build());

// Use Claude
genkit.generate(GenerateOptions.builder()
    .model("anthropic/claude-sonnet-4-5-20250929")
    .prompt("Hello!").build());

// Use a local Ollama model
genkit.generate(GenerateOptions.builder()
    .model("ollama/gemma3n")
    .prompt("Hello!").build());

Providing context documents

Pass documents for context-aware generation (useful for RAG):

List<Document> docs = List.of(
    Document.fromText("Paris is the capital of France."),
    Document.fromText("Berlin is the capital of Germany.")
);

ModelResponse response = genkit.generate(
    GenerateOptions.builder()
        .model("openai/gpt-4o-mini")
        .prompt("What is the capital of France?")
        .docs(docs)
        .build());

Using tools

Pass tools that models can call during generation:

ModelResponse response = genkit.generate(
    GenerateOptions.builder()
        .model("openai/gpt-4o")
        .prompt("What's the weather in Paris?")
        .tools(List.of(weatherTool))
        .build());

See Tool Calling for complete documentation on defining and using tools.

Next steps

Structured Output Generate type-safe outputs

Streaming Stream responses in real-time

Creating Flows Wrap generation in observable workflows