Implementation:Langchain ai Langchain BaseChatOpenAI With Structured Output

Knowledge Sources	LangChain OpenAI Structured Outputs
Domains	NLP, Data_Extraction, Structured_Output
Last Updated	2026-02-11 00:00 GMT

Overview

Concrete tool for extracting structured data from LLM responses provided by the LangChain OpenAI integration.

Description

The BaseChatOpenAI.with_structured_output() method returns a Runnable that chains the model with an output parser. It supports three extraction methods: function_calling (uses bind_tools + PydanticToolsParser), json_mode (forces JSON output + JsonOutputParser), and json_schema (OpenAI Structured Outputs with schema enforcement). The ChatOpenAI subclass defaults to json_schema method.

Usage

Call with_structured_output() with a Pydantic model or JSON schema to get type-safe responses from the model.

Code Reference

Source Location

• Repository: langchain
• File: libs/partners/openai/langchain_openai/chat_models/base.py
• Lines: L1961-2078 (BaseChatOpenAI), L3080-3170 (ChatOpenAI override)

Signature

def with_structured_output(
    self,
    schema: _DictOrPydanticClass | None = None,
    *,
    method: Literal[
        "function_calling", "json_mode", "json_schema"
    ] = "function_calling",
    include_raw: bool = False,
    strict: bool | None = None,
    tools: list | None = None,
    **kwargs: Any,
) -> Runnable[LanguageModelInput, _DictOrPydantic]:

Import

from langchain_openai import ChatOpenAI
from pydantic import BaseModel

I/O Contract

Inputs

Name	Type	Required	Description
schema	type[BaseModel] or dict or None	Yes	Pydantic model or JSON schema for output
method	str	No (default: "function_calling")	Extraction method
include_raw	bool	No (default: False)	Include raw AIMessage alongside parsed output
strict	bool or None	No	Enforce strict schema matching

Outputs

Name	Type	Description
return	Runnable	Returns Pydantic instance or dict matching schema; with include_raw=True returns {"raw": AIMessage, "parsed": ..., "parsing_error": ...}

Usage Examples

Pydantic Schema

from langchain_openai import ChatOpenAI
from pydantic import BaseModel, Field

class Person(BaseModel):
    name: str = Field(description="Person's full name")
    age: int = Field(description="Person's age")
    occupation: str = Field(description="Person's job title")

llm = ChatOpenAI(model="gpt-4o-mini")
structured_llm = llm.with_structured_output(Person)

result = structured_llm.invoke("Tell me about Marie Curie")
print(type(result))  # <class 'Person'>
print(result.name)   # "Marie Curie"
print(result.age)    # 66

With Raw Output

structured_llm = llm.with_structured_output(Person, include_raw=True)
result = structured_llm.invoke("Tell me about Einstein")
print(result["parsed"].name)  # "Albert Einstein"
print(result["raw"].content)  # Raw AIMessage
print(result["parsing_error"])  # None if successful

Related Pages

Implements Principle

• Principle:Langchain_ai_Langchain_Structured_Output_Extraction
• Environment:Langchain_ai_Langchain_OpenAI_API_Credentials