fix(core): use getBufferString for message summarization

[christian-bromann]

Port of langchain-ai/langchain#34607

Problem

Fixes token inflation in SummarizationMiddleware that caused context window overflow during summarization.

Root cause: When formatting messages for the summary prompt, JSON.stringify(messages) was being used, which includes all metadata fields (usage_metadata, response_metadata, additional_kwargs, etc.). This caused the stringified representation to use ~2.5x more tokens than countTokensApproximately estimates.

Symptoms:

• Summarization triggers at expected threshold based on countTokensApproximately
• But JSON.stringify(messages) in the prompt uses significantly more tokens
• Can result in context length exceeded errors

Solution

Use getBufferString() to format messages, which produces compact output:

Human: What's the weather?
AI: Let me check...[tool_calls]
Tool: 72°F and sunny

Instead of verbose JSON representation:

[
  {
    "type": "human",
    "content": "What's the weather?",
    "additional_kwargs": {},
    "response_metadata": {}
  },
  ...
]

Changes

• @langchain/core: Updated getBufferString() to use message's text property for compact content extraction, and added tool_calls/function_call support for AI messages
• langchain: Updated createSummary() in summarization middleware to use getBufferString() instead of JSON.stringify()
• Updated tests to reflect new expected format

[changeset-bot]

🦋 Changeset detected

Latest commit: 894a78a

The changes in this PR will be included in the next version bump.

This PR includes changesets to release 11 packages

Name	Type
@langchain/core	Patch
langchain	Patch
@langchain/standard-tests	Patch
@langchain/anthropic	Patch
@langchain/google-cloud-sql-pg	Patch
@langchain/google-common	Patch
@langchain/google-genai	Patch
@langchain/google-webauth	Patch
@langchain/google-vertexai-web	Patch
@langchain/google-vertexai	Patch
@langchain/google-gauth	Patch

Not sure what this means? Click here to learn what changesets are.

Click here if you're a maintainer who wants to add another changeset to this PR