Skip to main content
Use PydanticAI when you want Python agents with typed outputs, validation, and a compact OpenAI-compatible model setup.

​
Install

pip install "pydantic-ai-slim[openai]"

​
Configure the model

import os
from pydantic_ai import Agent
from pydantic_ai.models.openai import OpenAIChatModel
from pydantic_ai.providers.openai import OpenAIProvider

model = OpenAIChatModel(
    "openai/gpt-5-nano",
    provider=OpenAIProvider(
        api_key=os.environ["AI_STATS_API_KEY"],
        base_url="https://api.phaseo.app/v1",
    ),
)

agent = Agent(model, system_prompt="Answer concisely.")
result = agent.run_sync("Reply with only: ok")
print(result.output)

​
Notes

  • PydanticAI’s OpenAI-compatible path uses OpenAI-style chat behavior.
  • Use direct AI Stats Responses calls for Responses-only features such as ai-stats:apply_patch.
  • Keep structured output schemas small enough for the selected model’s context window.
Last modified on June 11, 2026