Topics / #testing-tool

Testing-tool projects on GitHub

End-to-end and automated testing tools for apps and APIs.

As of , the most-starred Testing-tool projects on GitHub are cypress (51K stars), testify (26.2K stars) and keploy (18.6K stars). GitHub.Locker tracks 7 Testing-tool repositories with 140.3K stars combined. The fastest-growing on October 16 was artemis, with +143 stars in a day.

7Projects
140.3KStars in total
+174Stars in a day

Hot right now

Most stars gained in a day · Oct 16

The all-time top

7 projects · by stars

  1. #1 · cypress-io/cypress

    cypress

    Cypress is a JavaScript/TypeScript testing framework for anything that runs in a browser, built to make end-to-end and component testing fast, easy, and reliable.

    51K
  2. #2 · stretchr/testify

    testify

    Testify is a toolkit of Go packages that add assertions, mocking and test suites on top of the standard library, so tests are easier to write and failures are easy to read.

    26.2K
  3. #3 · keploy/keploy

    keploy

    Keploy is an open-source tool that auto-generates API and integration tests from real traffic: it records your API calls, database queries and streaming events, then replays them as tests.

    18.6K
  1. 4

    jasmine

    15.8Kjasmine/jasmine · JavaScript

    Jasmine is a Behavior Driven Development (BDD) testing framework for JavaScript, used to write and run specs (tests) for your code.

  2. 5

    headless-recorder

    15.3Kcheckly/headless-recorder · JavaScript

    Headless Recorder is a Chrome extension that records your clicks and typing in the browser and generates a ready-to-run Playwright or Puppeteer script from them.

  3. 6

    artemis

    12.1Kgoogle/artemis · Python

    ARTEMIS lets AI assistants and test suites control real Android phones using natural-language instructions, automating end-to-end workflows while capturing logs and screenshots along the way.

  4. 7

    reticle

    1.2Kreticlehq/reticle · TypeScript

    Reticle lets a coding AI agent verify its own work by driving the real running app and reading what actually happened — network calls, app state, console — instead of just assuming the code works, returning a pass, fail, or 'unknown' verdict with the exact file:line to fix.

Often tagged together