Reducers and Annotated

A reducer decides how an old state value and a new update become one value. Without a reducer, a new update replaces the old value.¹

Tiny example

import operator
from typing import Annotated
from typing_extensions import TypedDict

class ResearchState(TypedDict):
    question: str
    findings: Annotated[list[str], operator.add]

def search_policy(state: ResearchState) -> dict:
    return {"findings": ["Refund window is 30 days"]}

Annotated[list[str], operator.add] means “append this update to the existing list.”

Update	No reducer	`operator.add` reducer
old `['A']`, new `['B']`	`['B']`	`['A', 'B']`

Failure note

Parallel nodes that update the same key need a merge rule. Choose a deterministic reducer and do not assume parallel results arrive in a business-significant order.

LangChain, Use the graph API. ↩