PydanticAI

PydanticAI is worth adding because it brings a very distinctive value proposition to the agent-framework space: Python agents built around typed models, validation, and structured outputs using the same ergonomics that made Pydantic broadly successful. The official docs position it as a GenAI agent framework, and the product shape is clearly about reliable Python development rather than abstract multi-agent theory or SaaS platform packaging.

Its pricing posture is open-source. The official docs and repository focus on the framework, installation, and provider integrations rather than any direct subscription. That means PydanticAI is free to adopt at the software layer, while real cost comes from the models, observability, and infrastructure you pair with it. For teams already comfortable with Pydantic-style development, that can make PydanticAI one of the cleanest ways to bring typed discipline into agent systems without buying a whole platform first.