zod

To validate data against a Zod schema, use safeParse for synchronous parsing and parseAsync for async refinements, which returns structured error handling with paths and flatten outputs for comprehensive issue reporting.

The best way to ensure type-safe schema composition in Zod is to use z.infer for type derivation, discriminated unions for reliable type narrowing, and pick/omit variants for partial updates, avoiding z.any.

You need to use parseAsync instead of safeParse for TypeScript validation when your schema includes asynchronous refinements, ensuring async operations are properly resolved during the parsing strategy.

Yes, Zod supports internationalization for error reporting by providing path-based and i18n-ready error handling outputs, allowing you to map validation errors to localized messages across frontend and backend validation.

Performance optimizations for parsing large Zod schemas include schema caching, lazy loading techniques, and considering Zod Mini to reduce runtime overhead during type-safe validation and parsing.

You should avoid z.any in Zod schema design because it breaks type safety, preventing the compiler from catching runtime validation bugs and undermining robust error handling and reliable type narrowing.