Structured output not working with system prompt

[jewells77]

Checked other resources

• This is a bug, not a usage question.
• I added a clear and descriptive title that summarizes this issue.
• I used the GitHub search to find a similar question and didn't find it.
• I am sure that this is a bug in LangChain rather than my code.
• The bug is not resolved by updating to the latest stable version of LangChain (or the specific integration package).
• This is not related to the langchain-community package.
• I read what a minimal reproducible example is (https://stackoverflow.com/help/minimal-reproducible-example).
• I posted a self-contained, minimal, reproducible example. A maintainer can copy it and run it AS IS.

Example Code

from pydantic import BaseModel
from langchain.agents import create_agent
from langchain.agents.structured_output import ToolStrategy


class ContactInfo(BaseModel):
    name: str
    email: str
    phone: str

agent = create_agent(
    model="openai:gpt-4o-mini",
    tools=[search_tool],
    response_format=ToolStrategy(ContactInfo)
)

result = agent.invoke({
    "messages": [{"role": "user", "content": "Extract contact info from: John Doe, john@example.com, (555) 123-4567"}]
})

result["structured_response"]
# ContactInfo(name='John Doe', email='john@example.com', phone='(555) 123-4567')

When we include system_prompt it causes an error.

agent = create_agent(
    model="openai:gpt-4o-mini",
    tools=[search_tool],
    system_prompt='you are an expert in extracting data',
    response_format=ToolStrategy(ContactInfo)
)

Error Message and Stack Trace (if applicable)

Error processing chat: Error code: 400 - {'error': {'message': "Invalid type for 'messages[0].content[0]': expected an object, but got a string instead.", 'type': 'invalid_request_error', 'param': 'messages[0].content[0]', 'code': 'invalid_type'}}

Description

What I am doing:
Creating an agent with create_agent from langchain.agents that uses a structured response format (ToolStrategy(ContactInfo)) and a model (openai:gpt-4o-mini). The goal is to extract contact information from text.

What I expect:
The agent should initialize successfully and process the input message, returning a structured ContactInfo object as it does without the system_prompt.

System Info

System Information

OS: Windows
OS Version: 10.0.26200
Python Version: 3.12.10 (main, Apr 9 2025, 04:06:22) [MSC v.1943 64 bit (AMD64)]

Package Information

langchain_core: 1.0.1
langchain: 1.0.2
langsmith: 0.4.38
langchain_google_genai: 3.0.0
langchain_mongodb: 0.7.1
langchain_openai: 1.0.1
langchain_text_splitters: 1.0.0
langgraph_sdk: 0.2.9

Optional packages not installed

langserve

Other Dependencies

claude-agent-sdk: Installed. No version info available.
filetype: 1.2.0
google-ai-generativelanguage: 0.9.0
httpx: 0.28.1
jsonpatch: 1.33
langchain-anthropic: Installed. No version info available.
langchain-aws: Installed. No version info available.
langchain-community: Installed. No version info available.
langchain-deepseek: Installed. No version info available.
langchain-fireworks: Installed. No version info available.
langchain-google-vertexai: Installed. No version info available.
langchain-groq: Installed. No version info available.
langchain-huggingface: Installed. No version info available.
langchain-mistralai: Installed. No version info available.
langchain-ollama: Installed. No version info available.
langchain-perplexity: Installed. No version info available.
langchain-together: Installed. No version info available.
langchain-xai: Installed. No version info available.
langgraph: 1.0.1
langsmith-pyo3: Installed. No version info available.
lark: 1.3.0
numpy: 2.3.4
openai: 2.6.1
openai-agents: Installed. No version info available.
opentelemetry-api: Installed. No version info available.
opentelemetry-exporter-otlp-proto-http: Installed. No version info available.
opentelemetry-sdk: Installed. No version info available.
orjson: 3.11.4
packaging: 25.0
pydantic: 2.12.3
pymongo: 4.15.3
pytest: Installed. No version info available.
pyyaml: 6.0.3
requests: 2.32.5
requests-toolbelt: 1.0.0
rich: Installed. No version info available.
tenacity: 9.1.2
tiktoken: 0.12.0
typing-extensions: 4.15.0
vcrpy: Installed. No version info available.
zstandard: 0.25.0

[ccurme]

I'm unable to reproduce this using your code:

from pydantic import BaseModel
from langchain.agents import create_agent
from langchain.agents.structured_output import ToolStrategy


class ContactInfo(BaseModel):
    name: str
    email: str
    phone: str


def search_tool(query: str) -> str:
    """Run a search."""
    pass 


agent = create_agent(
    model="openai:gpt-4o-mini",
    tools=[search_tool],
    system_prompt='you are an expert in extracting data',
    response_format=ToolStrategy(ContactInfo)
)

result = agent.invoke({
    "messages": [{"role": "user", "content": "Extract contact info from: John Doe, john@example.com, (555) 123-4567"}]
})

result["structured_response"]
# ContactInfo(name='John Doe', email='john@example.com', phone='(555) 123-4567')

would you mind sharing the entire traceback? thanks.