pydantic-jwt library brings type-safe JWT handling to Python with Pydantic models

A developer has released pydantic-jwt, a Python library that replaces untyped dictionary-based JWT payloads with structured Pydantic models, offering full type safety and editor autocompletion. The library allows a single model class to handle both token generation and parsing, while automatically validating standard claims such as expiry, issuer, and audience. It integrates cleanly with FastAPI's dependency injection system, enabling endpoints to receive verified, typed token objects instead of raw strings or dicts. Version 1.0.0 introduced a verified_only configuration flag that closes a security gap where FastAPI could deserialize a plain JSON body directly into a token model without any signature verification. The author openly documented the earlier vulnerability and the iterative attempts required to fix it safely within Pydantic's validation internals.