From 247569795588919345136152276702b6265c4e8d Mon Sep 17 00:00:00 2001 From: filipi87 Date: Tue, 10 Feb 2026 18:59:12 -0300 Subject: [PATCH] Changelog entries for context summarization --- changelog/3621.added.2.md | 1 + changelog/3621.added.md | 5 +++++ 2 files changed, 6 insertions(+) create mode 100644 changelog/3621.added.2.md create mode 100644 changelog/3621.added.md diff --git a/changelog/3621.added.2.md b/changelog/3621.added.2.md new file mode 100644 index 000000000..395c8bdfb --- /dev/null +++ b/changelog/3621.added.2.md @@ -0,0 +1 @@ +- Added new frames for context summarization: `LLMContextSummaryRequestFrame` and `LLMContextSummaryResultFrame`. diff --git a/changelog/3621.added.md b/changelog/3621.added.md new file mode 100644 index 000000000..8b2197eea --- /dev/null +++ b/changelog/3621.added.md @@ -0,0 +1,5 @@ +- Added context summarization feature to automatically compress conversation history when conversation length limits (by token or message count) are reached, enabling efficient long-running conversations. + - Configure via `enable_context_summarization=True` in `LLMAssistantAggregatorParams` + - Customize behavior with `LLMContextSummarizationConfig` (max tokens, thresholds, etc.) + - Automatically preserves incomplete function call sequences during summarization + - See new examples: `examples/foundational/54-context-summarization-openai.py` and `examples/foundational/54a-context-summarization-google.py`