Type Stubs for PyJWT

1.7.1 · deprecated · verified Thu Apr 16

types-pyjwt is a PEP 561 compliant type stub package providing static type annotations for the PyJWT library. It enables type-checking tools like MyPy, Pyright, and PyCharm to perform static analysis on code using PyJWT. The package itself is auto-generated from the community-maintained typeshed repository. As of PyPI version 1.7.1, it has not seen updates since mid-2021, as the upstream PyJWT library started including its own type annotations from version 2.0.0 onwards.

Common errors

```
error: Duplicate module named 'jwt' (or similar type conflict errors)
```
cause Using `types-pyjwt` simultaneously with `PyJWT` version 2.0.0 or newer, which ships with its own type annotations. Type checkers detect two sources of type information for the same module.

fix Uninstall `types-pyjwt` if your `PyJWT` version is 2.0.0 or higher: `pip uninstall types-pyjwt`.
```
error: No type hints found for package 'jwt' (or 'module 'jwt' has no attribute 'encode' (or 'decode')' with type checker)
```
cause Attempting to type-check `PyJWT` code without any type information. This occurs if `PyJWT < 2.0.0` is installed and `types-pyjwt` is missing, or if `types-pyjwt` is installed but the type checker isn't configured correctly to find stubs.

fix For `PyJWT < 2.0.0`, install `types-pyjwt`: `pip install types-pyjwt`. For `PyJWT >= 2.0.0`, ensure `types-pyjwt` is *not* installed and that your type checker recognizes `py.typed` packages (most modern type checkers do by default).

Warnings

breaking The `PyJWT` library itself includes type annotations starting from version 2.0.0. If you are using `PyJWT` version 2.0.0 or newer, you MUST uninstall `types-pyjwt` to avoid type conflicts, incorrect type checking results, or outdated stub definitions.
Fix: Uninstall `types-pyjwt`: `pip uninstall types-pyjwt`
deprecated `types-pyjwt` has not been updated since mid-2021 (version 1.7.1) and is effectively deprecated due to `PyJWT`'s self-typing. Further type improvements or fixes for `PyJWT` are now made directly within the `PyJWT` project.
Fix: Upgrade to `PyJWT` version 2.0.0 or newer and uninstall `types-pyjwt`.
gotcha When `types-pyjwt` is installed, type checkers may prefer its stubs over inline annotations in `PyJWT < 2.0.0`. However, for `PyJWT >= 2.0.0`, the presence of `types-pyjwt` can cause confusing errors or `mypy` to complain about duplicate definitions.
Fix: Consistently use either `types-pyjwt` (for `PyJWT < 2.0.0`) or PyJWT's built-in types (for `PyJWT >= 2.0.0`). The recommended approach is to upgrade `PyJWT` and remove `types-pyjwt`.

Install

pip install types-pyjwt Install types-pyjwt

Imports

jwt
```
import jwt
from jwt import PyJWT, decode, encode
```
types-pyjwt does not provide runtime objects to import directly. It provides type hints for the `jwt` module from the actual `PyJWT` library. Imports should always be from `jwt`.

Quickstart

This quickstart demonstrates a basic usage of PyJWT with type hints. When types-pyjwt is installed alongside PyJWT (specifically PyJWT versions prior to 2.0.0), a static type checker would use the stubs from `types-pyjwt` to validate the types in this code. For PyJWT versions 2.0.0 and newer, PyJWT includes its own `py.typed` file and inline annotations, making `types-pyjwt` unnecessary and potentially problematic.

import jwt
from typing import Dict, Any

def create_and_decode_token(payload: Dict[str, Any], secret: str) -> Dict[str, Any]:
    # This example assumes PyJWT < 2.0.0 would use types-pyjwt for type hints.
    # For PyJWT >= 2.0.0, the types are bundled.
    encoded_jwt = jwt.encode(payload, secret, algorithm="HS256")
    decoded_payload = jwt.decode(encoded_jwt, secret, algorithms=["HS256"])
    return decoded_payload

if __name__ == "__main__":
    my_payload = {"user_id": 123, "name": "test_user"}
    jwt_secret = "super-secret-key"
    
    # Run a type checker like MyPy on this file to see types-pyjwt in action (if PyJWT < 2.0.0)
    # Or, if PyJWT >= 2.0.0, to see PyJWT's bundled types.
    result = create_and_decode_token(my_payload, jwt_secret)
    print(f"Decoded payload: {result}")
    # Expected output for type checking (e.g., with mypy): No issues if types are correctly applied.

view raw JSON →