Typing stubs for PyJWT (jwt package)

0.1.3 · active · verified Thu Apr 16

This library provides PEP 561 type stubs for the popular `PyJWT` library, which is commonly imported as `jwt`. It enables static type checkers like MyPy, PyCharm, and Pyright to analyze code that uses `PyJWT`. `types-jwt` is part of the `typeshed` project, which centrally manages type definitions for many third-party Python packages. `types-jwt` version 0.1.3 was released in May 2021. However, it's important to note that since `PyJWT` version 2.0.0 (released December 2020) and later include their own inline type annotations, `types-jwt` is primarily relevant for projects using older versions of `PyJWT` (pre-2.0.0) or specific legacy setups.

Common errors

```
ModuleNotFoundError: No module named 'jwt'
```
cause The `PyJWT` runtime library, which provides the `jwt` module, is not installed in your Python environment. You may have only installed `types-jwt`, which is a stub-only package.

fix Install the `PyJWT` library: `pip install PyJWT`.
```
MyPy error: Incompatible types in assignment (expression has type "bytes", variable has type "str") or similar type errors related to `jwt.encode()` or `jwt.decode()` when both `PyJWT >= 2.0.0` and `types-jwt` are installed.
```
cause There is a conflict or redundancy between the inline type annotations provided by `PyJWT` version 2.0.0+ and the external stubs from `types-jwt`. `PyJWT` itself began including type information in version 2.0.0.

fix Uninstall `types-jwt` if you are using `PyJWT` version 2.0.0 or newer: `pip uninstall types-jwt`. Rely on the inline types from `PyJWT`.
```
MyPy error: 'Module' object has no attribute 'encode' or 'decode' even after installing types-jwt.
```
cause This often occurs if `PyJWT` is not installed, or if there's an environment issue where MyPy cannot find the correct `PyJWT` installation to apply the stubs to. Also, older `PyJWT` versions might have inconsistent typing.

fix Ensure `PyJWT` is installed (`pip install PyJWT`). Verify your environment setup. If using `PyJWT < 2.0.0`, ensure `types-jwt` is installed. If using `PyJWT >= 2.0.0`, ensure `types-jwt` is *not* installed and that `PyJWT`'s own `py.typed` file is being recognized.

Warnings

breaking `types-jwt` (and `types-PyJWT`) are largely redundant for `PyJWT` versions 2.0.0 and higher. Starting with `PyJWT` 2.0.0 (released December 2020), the `PyJWT` library itself includes inline type annotations. Users should uninstall `types-jwt` if they are using `PyJWT >= 2.0.0` to avoid potential conflicts, redundant packages, or unexpected type-checking behavior.
Fix: If using `PyJWT >= 2.0.0`, uninstall `types-jwt`: `pip uninstall types-jwt`. Rely on `PyJWT`'s built-in type hints.
gotcha Installing `types-jwt` does NOT install the actual `PyJWT` runtime library. `types-jwt` only provides static type hints. To use JWT functionality (e.g., `jwt.encode`, `jwt.decode`), you must explicitly install `PyJWT` (e.g., `pip install PyJWT`) in addition to `types-jwt`.
Fix: Ensure both `types-jwt` (for type checking) and `PyJWT` (for runtime functionality) are installed: `pip install types-jwt PyJWT`.
gotcha Type stub packages from `typeshed` are designed for static analysis tools (like MyPy) and should not be directly imported in your application's runtime code. All imports for JWT operations should come from the actual `PyJWT` library (i.e., `import jwt` or `from jwt import ...`). Importing from `types_jwt` will result in a `ModuleNotFoundError` or `ImportError` at runtime.
Fix: Always import from `jwt` (the runtime library) and *never* from `types_jwt` in your Python code: `import jwt`.

Install

pip install types-jwt Install types-jwt

Imports

encode
wrong:
```
from types_jwt import encode
```
correct:
```
from jwt import encode
```
`types-jwt` is a stub-only package. You import the runtime functions from the actual `PyJWT` library (which uses the `jwt` module name).
decode
wrong:
```
from types_jwt import decode
```
correct:
```
from jwt import decode
```
`types-jwt` is a stub-only package. You import the runtime functions from the actual `PyJWT` library (which uses the `jwt` module name).

Quickstart

This quickstart demonstrates how to encode and decode a JSON Web Token using the `PyJWT` library, for which `types-jwt` provides type annotations. The example includes setting an expiration time and handling basic decoding errors. Ensure you have `PyJWT` installed for this code to run. While `types-jwt` itself is not directly imported, its presence allows type checkers to provide richer feedback on `jwt` module usage. For production, always use robust, securely generated secret keys and appropriate algorithms.

import jwt
import os
from datetime import datetime, timedelta, timezone
from typing import Dict, Any # These types are enhanced by types-jwt if PyJWT < 2.0

# Note: This quickstart uses PyJWT (the runtime library)
# types-jwt provides type hints for this code during static analysis.

# Replace with a strong, secret key in a real application
secret_key: bytes = os.environ.get('JWT_SECRET_KEY', 'your-super-secret-key').encode('utf-8')
algorithm: str = "HS256"

# 1. Encode a JWT
payload: Dict[str, Any] = {
    "user_id": 123,
    "username": "testuser",
    "exp": datetime.now(timezone.utc) + timedelta(hours=1), # Token expires in 1 hour
    "iat": datetime.now(timezone.utc) # Issued at time
}

try:
    encoded_jwt: str = jwt.encode(payload, secret_key, algorithm=algorithm)
    print(f"Encoded JWT: {encoded_jwt}")

    # 2. Decode the JWT
    decoded_payload: Dict[str, Any] = jwt.decode(
        encoded_jwt, secret_key, algorithms=[algorithm]
    )
    print(f"Decoded Payload: {decoded_payload}")

except jwt.ExpiredSignatureError:
    print("Token has expired!")
except jwt.InvalidTokenError as e:
    print(f"Invalid token: {e}")

view raw JSON →