Add Python Type Hints to Legacy Code with Claude

Legacy Python code without type hints is hard to refactor safely and painful to navigate in an IDE. Paste a function or module and ask Claude to add proper type annotations — it infers types from usage patterns, docstrings, and naming conventions, and handles complex cases like Optional, Union, and generics correctly.

# Before — no type hints, hard to refactor safely
def process_orders(orders, discount_rate=None):
    results = []
    for order in orders:
        total = order['amount'] * (1 - (discount_rate or 0))
        results.append({'id': order['id'], 'total': total})
    return results

Claude annotates it and suggests stronger types:

from typing import Optional, TypedDict

class Order(TypedDict):
    id: int
    amount: float

class OrderResult(TypedDict):
    id: int
    total: float

def process_orders(
    orders: list[Order],
    discount_rate: Optional[float] = None,
) -> list[OrderResult]:
    results: list[OrderResult] = []
    for order in orders:
        total = order['amount'] * (1 - (discount_rate or 0))
        results.append({'id': order['id'], 'total': total})
    return results

Ask Claude to replace loosely typed dict structures with TypedDict or dataclasses — this gives you better IDE autocompletion and makes the code safe to refactor. You can also ask it to generate a mypy config alongside the annotated code.

Type hints aren't just documentation — they're the foundation for confident refactoring.