Pydantic AI

Pydantic AI brings Pydantic’s type-driven validation to agent and tool design: typed dependencies, structured results, and model-agnostic runs. Tool arguments and outputs become schema-checked rather than “almost JSON.” For an agentic OS this is a strong fit for Python workers that must emit ledger-safe structured events and validated proposals before any gated side effect. It competes with Instructor/Mirascope in the typed-agent band. How Python does AI. Agents, realtime voice, image generation, embeddings. Every model, every interface, typed end to end. Agents, realtime voice, image generation, embeddings. Every model, every interface, typed end to end. Pydantic AI is the Python AI SDK: a typed, extensible agent loop with every model a string swap away. The same agent runs everywhere you need it: behind a web frontend, in the terminal, on a voice call, on a durable.