Skip to content

The agent engineering platform.

docs: add example of simultaneous tool-calling + structured output for OpenAI (#31433)

ccurmebbb60e210a09
diff --git a/docs/docs/integrations/chat/openai.ipynb b/docs/docs/integrations/chat/openai.ipynb
index 8df8e6d0c1..58bec4b2f6 100644
--- a/docs/docs/integrations/chat/openai.ipynb
+++ b/docs/docs/integrations/chat/openai.ipynb
@@ -397,6 +397,56 @@
"For more on binding tools and tool call outputs, head to the [tool calling](/docs/how_to/function_calling) docs."
]
},
+ {
+ "cell_type": "markdown",
+ "id": "f06789fb-61e1-4b35-a2b5-2dea18c1a949",
+ "metadata": {},
+ "source": [
+ "### Structured output and tool calls\n",
+ "\n",
+ "OpenAI's [structured output](https://platform.openai.com/docs/guides/structured-outputs) feature can be used simultaneously with tool-calling. The model will either generate tool calls or a response adhering to a desired schema. See example below:"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "id": "15d2b6e0-f457-4abd-a4d5-08b210d09c04",
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "from langchain_openai import ChatOpenAI\n",
+ "from pydantic import BaseModel\n",
+ "\n",
+ "\n",
+ "def get_weather(location: str) -> None:\n",
+ " \"\"\"Get weather at a location.\"\"\"\n",
+ " return \"It's sunny.\"\n",
+ "\n",
+ "\n",
+ "class OutputSchema(BaseModel):\n",
+ " \"\"\"Schema for response.\"\"\"\n",
+ "\n",
+ " answer: str\n",
+ " justification: str\n",
+ "\n",
+ "\n",
+ "llm = ChatOpenAI(model=\"gpt-4.1\")\n",
+ "\n",
+ "structured_llm = llm.bind_tools(\n",
+ " [get_weather],\n",
+ " response_format=OutputSchema,\n",
+ " strict=True,\n",
+ ")\n",
+ "\n",
+ "# Response contains tool calls:\n",
+ "tool_call_response = structured_llm.invoke(\"What is the weather in SF?\")\n",
+ "\n",
+ "# structured_response.additional_kwargs[\"parsed\"] contains parsed output\n",
+ "structured_response = structured_llm.invoke(\n",
+ " \"What weighs more, a pound of feathers or a pound of gold?\"\n",
+ ")"
+ ]
+ },
{
"cell_type": "markdown",
"id": "84833dd0-17e9-4269-82ed-550639d65751",