Pydantic Data Validation

Pydantic is the most widely used data validation library for Python, powered by type hints and built on Rust for performance. Created by Samuel Colvin in 2017, Pydantic has over 23,000 stars as of 2026 and is used by FastAPI, OpenAI, Uber, the NSA, and MongoDB. The V2 release in 2023 rewrote the core validation engine in Rust (via pydantic-core), achieving 5-50x performance improvements. Key features include: type hint-based validation (using Python type annotations for automatic field validation, supporting int, float, str, bool, datetime, UUID, URL, EmailStr, and custom types), BaseModel (the core class for structured data models with field types, defaults, validators, and computed fields), nested models (composing complex structures from nested models, lists, dicts, and optional fields), custom validators (field_validator and model_validator decorators for custom validation), serialization (model_dump and model_dump_json with field filtering), JSON schema generation (automatic JSON Schema from model definitions for API documentation), strict mode (enforcing strict type checking without coercion), dataclasses (validation for standard Python dataclasses), Settings management (BaseSettings for loading config from environment variables, .env files, and secrets directories), TypeAdapter (validating non-model types including generics, Union, and Literal), and Annotated types (PEP 593 for attaching metadata and constraints).