Ty is Astral’s fast Python type checker and language server, designed to provide diagnostics and editor intelligence with low feedback latency. It is part of a broader shift toward faster, integrated Python tooling.
Start by adding type hints to a small module rather than converting an entire codebase at once:Copy
def total(prices: list[float], tax: float = 0.0) -> float:
return sum(prices) * (1 + tax)A type checker catches mismatched assumptions before runtime, but it does not replace tests. Begin with function boundaries, configuration objects, and external data. Use explicit types where they improve communication; do not annotate every temporary variable just to increase a number.
Evaluate Ty against your project’s Python version, framework support, CI needs, editor setup, and existing checker configuration. Tooling can change quickly, especially while a project is developing. Pin versions in CI and review diagnostics rather than enabling an aggressive mode without a migration plan.
Whether you choose Ty or another checker, the habit matters most: make contracts visible, run checks automatically, and treat type errors as feedback about design – not merely obstacles to suppress.