Topics / #types

Types projects on GitHub

Static typing tools and type utility libraries for safer code.

As of , the most-starred Types projects on GitHub are nodebestpractices (105.7K stars), DefinitelyTyped (51.4K stars) and react (47.1K stars). GitHub.Locker tracks 6 Types repositories with 243.3K stars combined. The fastest-growing on October 16 was nodebestpractices, with +6 stars in a day.

6Projects
243.3KStars in total
+19Stars in a day

Hot right now

Most stars gained in a day · Oct 16

The all-time top

6 projects · by stars

  1. #1 · goldbergyoni/nodebestpractices

    nodebestpractices

    A curated, community-maintained list that summarizes the top Node.js best practices, style guides, and architectural tips, organized into chapters covering architecture, error handling, code style, testing, production readiness, security, performance, and Docker.

    105.7K
  2. #2 · DefinitelyTyped/DefinitelyTyped

    DefinitelyTyped

    DefinitelyTyped is a massive community repository that hosts TypeScript type definitions (.d.ts files) for JavaScript packages that don't ship their own types, published to npm under the @types scope.

    51.4K
  3. #3 · typescript-cheatsheets/react

    react

    A maintained cheatsheet for experienced React developers adopting TypeScript, packed with opinionated best practices and copy-pastable examples you can drop straight into your code.

    47.1K
  1. 4

    mypy

    20.7Kpython/mypy · Python

    Mypy is a static type checker for Python: you add type hints (PEP 484) to your code and it warns you when you use those types incorrectly, catching bugs without even running the program.

  2. 5

    type-fest

    17.5Ksindresorhus/type-fest · TypeScript

    type-fest is a collection of essential TypeScript types — helper types the author feels should have been built into the language, covering objects, unions, strings, arrays, tuples and functions.

  3. 6

    TypeLLM

    956TypeLLM/TypeLLM · Python

    TypeLLM adds type-safe, schema-guaranteed output generation to existing open autoregressive LLMs, without changing their weights or architecture, letting models still think freely while returning typed JSON Schema-constrained answers.

Often tagged together