Language Models

Language models provide simple text completion without conversation context, suitable for single-turn prompts and text generation tasks.

OllamaLanguageModel

Synchronous language model for blocking text completion.

Class Signature

package dev.langchain4j.model.ollama;

public class OllamaLanguageModel implements LanguageModel

Thread Safety: Immutable after build(); safe for concurrent requests

Nullability: Instance never null after successful build()

Public Methods

generate

public Response<String> generate(String prompt)

Generates text completion from a prompt.

Parameters:

• prompt - The input text prompt (must not be null or empty)

Returns: Response<String> containing generated text and token usage metadata

• Never null
• Content never null (may be empty string)
• Deterministic with same seed and temperature=0.0

Throws:

• IllegalArgumentException - If prompt is null or empty
• HttpTimeoutException - If request exceeds configured timeout
• IOException - If network connectivity fails
• RuntimeException - If Ollama server returns error

Thread Safety: Safe for concurrent calls

Retry Behavior: Automatically retries on transient failures up to maxRetries times

Example:

import dev.langchain4j.model.output.Response;
import java.io.IOException;

try {
    Response<String> response = model.generate("Once upon a time");
    String text = response.content();
    TokenUsage usage = response.tokenUsage();
    System.out.println("Generated: " + text);
} catch (IOException e) {
    System.err.println("Network error: " + e.getMessage());
}

OllamaStreamingLanguageModel

Streaming language model for real-time token-by-token text completion.

Class Signature

package dev.langchain4j.model.ollama;

public class OllamaStreamingLanguageModel implements StreamingLanguageModel

Thread Safety: Immutable after build(); safe for concurrent requests

Streaming Threading: Callbacks invoked on HTTP client thread

Public Methods

generate

public void generate(String prompt, StreamingResponseHandler<String> handler)

Generates text completion from a prompt with streaming output.

Parameters:

• prompt - The input text prompt (must not be null or empty)
• handler - Handler for streaming response callbacks (must not be null)

Returns: void - Method returns immediately; response arrives via callbacks

Throws:

• IllegalArgumentException - If prompt or handler is null

Error Handling: Errors during streaming trigger handler.onError(Throwable)

No Retry: Streaming operations do not automatically retry

Thread Safety: Handler must be thread-safe if shared across calls

See Also

• Chat Models - For conversational AI
• Embedding Model - For text embeddings
• Request Parameters - Detailed parameter documentation