Agent with Structured Output

Code

1from typing import List
2
3from kern.agent import Agent, RunOutput # noqa
4from kern.models.ibm import WatsonX
5from pydantic import BaseModel, Field
6from rich.pretty import pprint # noqa
7
8class MovieScript(BaseModel):
9 setting: str = Field(
10 ..., description="Provide a nice setting for a blockbuster movie."
11 )
12 ending: str = Field(
13 ...,
14 description="Ending of the movie. If not available, provide a happy ending.",
15 )
16 genre: str = Field(
17 ...,
18 description="Genre of the movie. If not available, select action, thriller or romantic comedy.",
19 )
20 name: str = Field(..., description="Give a name to this movie")
21 characters: List[str] = Field(..., description="Name of characters for this movie.")
22 storyline: str = Field(
23 ..., description="3 sentence storyline for the movie. Make it exciting!"
24 )
25
26movie_agent = Agent(
27 model=WatsonX(id="mistralai/mistral-small-3-1-24b-instruct-2503"),
28 description="You help people write movie scripts.",
29 output_schema=MovieScript,
30)
31
32# Get the response in a variable
33# movie_agent: RunOutput = movie_agent.run("New York")
34# pprint(movie_agent.content)
35
36movie_agent.print_response("New York")

Usage

Set up your virtual environment

1uv venv --python 3.12
2source .venv/bin/activate
1uv venv --python 3.12
2.venv\Scripts\activate

Set your API key

1export IBM_WATSONX_API_KEY=xxx
2export IBM_WATSONX_PROJECT_ID=xxx

Install dependencies

1uv pip install -U ibm-watsonx-ai pydantic rich kern-ai

Run Agent

1python cookbook/11_models/ibm/watsonx/structured_output.py

This example shows how to use structured output with IBM WatsonX. It defines a Pydantic model MovieScript with various fields and their descriptions, then creates an agent using this model as the output_schema. The model's output will be parsed into this structured format.