Skip to content

SummarizationStrategy

This docs was updated at: 2026-02-23

com.paragon.agents.context.SummarizationStrategy  ยท  Class

Implements ContextWindowStrategy

A context window strategy that summarizes older messages when context exceeds the limit.

This strategy:

  • Keeps recent messages intact
  • Summarizes older messages using an LLM
  • Replaces summarized messages with a single summary message
  • Caches summaries to avoid redundant API calls

This strategy is more expensive (requires an LLM call) but preserves more context than simple truncation, making it suitable for conversations where older context contains important information.

Usage Example

// Use agent's responder with a fast model for summarization
ContextWindowStrategy strategy = SummarizationStrategy.builder()
    .responder(responder)
    .model("openai/gpt-4o-mini")
    .build();
// Use with agent
Agent agent = Agent.builder()
    .name("Assistant")
    .model("openai/gpt-4o")
    .instructions("You are a helpful assistant.")
    .responder(responder)
    .contextWindow(strategy, 4000)
    .build();

See Also

  • ContextWindowStrategy
  • SlidingWindowStrategy

Since: 1.0

Methods

builder

public static @NonNull Builder builder()

Creates a new builder for SummarizationStrategy.

Returns

a new builder instance

withResponder

public static @NonNull SummarizationStrategy withResponder(
      @NonNull Responder responder, @NonNull String model)

Creates a summarization strategy with the given responder and model.

Parameters

Name Description
responder the responder to use for summarization calls
model the model to use for summarization

Returns

a new summarization strategy

generateSummary

private String generateSummary(List<ResponseInputItem> messages)

Generates a summary of the given messages using the configured responder.

formatMessage

private String formatMessage(ResponseInputItem item)

Formats a message for summarization.

model

public @NonNull String model()

Returns the model used for summarization.

Returns

the model identifier

keepRecentMessages

public int keepRecentMessages()

Returns the number of recent messages to keep.

Returns

the keep recent messages count

responder

public @NonNull Builder responder(@NonNull Responder responder)

Sets the responder to use for summarization calls.

Parameters

Name Description
responder the responder

Returns

this builder

model

public @NonNull Builder model(@NonNull String model)

Sets the model to use for summarization.

A fast, cheaper model is recommended (e.g., "openai/gpt-4o-mini").

Parameters

Name Description
model the model identifier

Returns

this builder

summarizationPrompt

public @NonNull Builder summarizationPrompt(@NonNull String prompt)

Sets a custom prompt for summarization.

The prompt should contain a single %s placeholder where the conversation text will be inserted.

Parameters

Name Description
prompt the summarization prompt

Returns

this builder

keepRecentMessages

public @NonNull Builder keepRecentMessages(int count)

Sets the number of recent messages to keep without summarization.

Defaults to 5 messages.

Parameters

Name Description
count the number of messages to keep

Returns

this builder

build

public @NonNull SummarizationStrategy build()

Builds the SummarizationStrategy.

Returns

a new SummarizationStrategy

Throws

Type Condition
NullPointerException if responder or model is null