Implementation:Langchain ai Langchain AIMessageChunk Accumulation

Knowledge Sources	LangChain
Domains	Streaming, Data_Aggregation
Last Updated	2026-02-11 00:00 GMT

Overview

Concrete tool for accumulating streaming chunks into a complete response using the AIMessageChunk addition operator provided by langchain-core.

Description

The AIMessageChunk.__add__() method merges two chunks by concatenating content strings, merging tool_call_chunks by index, combining usage_metadata, and merging response_metadata. This enables reconstructing the full response from a stream.

Usage

Use the + operator on AIMessageChunk objects to accumulate a full response from streaming output.

Code Reference

Source Location

• Repository: langchain
• File: libs/core/langchain_core/messages/ai.py
• Lines: L413-500+

Signature

class AIMessageChunk(AIMessage, BaseMessageChunk):
    type: Literal["AIMessageChunk"] = "AIMessageChunk"
    tool_call_chunks: list[ToolCallChunk] = Field(default_factory=list)

    def __add__(self, other: AIMessageChunk) -> AIMessageChunk:
        # Merges content, tool_call_chunks, usage_metadata, response_metadata
        ...

Import

from langchain_core.messages import AIMessageChunk

I/O Contract

Inputs

Name	Type	Required	Description
self	AIMessageChunk	Yes	Accumulated chunk so far
other	AIMessageChunk	Yes	New chunk to merge

Outputs

Name	Type	Description
return	AIMessageChunk	Merged chunk with combined content, tool calls, and metadata

Usage Examples

Accumulating a Full Response

from langchain_openai import ChatOpenAI

llm = ChatOpenAI(model="gpt-4o-mini")

full = None
for chunk in llm.stream("Explain Python in 3 sentences"):
    print(chunk.content, end="", flush=True)
    full = chunk if full is None else full + chunk

print(f"\n\nFull content: {full.content}")
print(f"Usage: {full.usage_metadata}")

Related Pages

Implements Principle

• Principle:Langchain_ai_Langchain_Chunk_Accumulation