AI testing and QA agents
Agents that turn requirements or plain-English steps into tests, run them in CI and repair them when the product changes.
By weekly package downloads, the leaders are e2e (1.6M), Momentic (939K). By GitHub stars, e2e leads (14.1K); the others are far behind. Among agents with at least 20,000 downloads in the last 30 days, the fastest grower is Momentic, with downloads up 112%. As of Nov 21, 2026.
-
23
56 939K up 112.0% — — 939K downloads a week up 112.0% 30dGid 56 Free + $500/mo
-
39
51 1.6M — 14.1K 2d ago 1.6M downloads a weekGid 51 Free (OSS)e2e · Testing · open-source agentic end-to-end testing framework
A TypeScript SDK, runner and CLI for end-to-end tests of web and mobile apps in which agent steps carry out natural-language tasks in real browsers, simulators or emulators; an agent step is recorded once a later check passes and replayed on the next run without model calls, handing back to the agent when the app changed.
-
82
41 12.6K up 7.6% — — 12.6K downloads a week up 7.6% 30dGid 41 Custom
-
157
312/4 19K up 30.5% — — 19K downloads a week up 30.5% 30dGid 31 · 2/4 signals Custom
-
244
16 1,925 up 68.6% 267 yesterday 1,925 downloads a week up 68.6% 30dGid 16 Free
-
255
52/4 — — — — Website rank #527K (Tranco top-sites list)Gid 5 · 2/4 signals Free + $1/moQA Wolf, Inc. · Testing · AI test automation platform and managed QA service
End-to-end testing platform whose AI agents map an application, generate Playwright (web) and Appium (mobile) tests, reproduce failures and repair broken tests, running them in parallel on hosted infrastructure for web, iOS, Android and Windows apps. Runs are triggered manually, on a schedule or from CI, and tests can be requested from Slack, GitHub or coding agents through an MCP server. A managed Coverage-as-a-Service option has QA engineers build and maintain the suite.
-
—
— — — — — Website rank #2M (Tranco top-sites list) $40/moIto Inc. · Testing · agentic QA for pull requests
AI QA agent for GitHub pull requests. On each PR it reads the diff and description, plans tests for the affected user flows, builds a single-use copy of the app in an isolated sandbox and tests the frontend and real backend with computer-use agents, without test scripts. It posts results to the PR, with video replays, logs, repro steps and severity for failures, and re-runs after fixes. It accepts test credentials, including multi-factor authentication.
-
—
— — — — — 2,436 VS Code installs —ElevenXN Technologies · Testing · IDE extension with SDLC and testing agents
IDE extension for VS Code, Cursor and Windsurf with AI agents for planning, coding, testing and deployment. Includes Plan and Agent modes plus API, web (Playwright/Cypress/Selenium), unit and mobile (Appium) test agents, MCP integrations with tools like Jira and GitHub, and support for 40+ model providers.
-
—
— — — — — Website rank #266K (Tranco top-sites list) FreetestRigor · Testing · generative-AI test automation (plain-English tests)
Generative-AI test automation tool in which tests are written in plain English and turned into UI steps that users can review. It runs end-to-end tests on web browsers, native and hybrid iOS and Android apps, Windows desktop apps (paid plans), APIs, email, SMS, phone calls, 2FA and mainframe apps, finding elements as they appear on screen rather than by locators. Integrates with CI/CD tools, Jira, TestRail and Claude Code via MCP.
-
—
— — — — — Website rank #173K (Tranco top-sites list) CustomTestsigma · Testing · multi-agent test automation platform
AI coworker in the Testsigma platform that uses specialised agents to plan, write, run, heal and analyse tests for web, mobile, API and Salesforce apps, with Windows desktop and SAP also covered. Agents map commits to stories, update locators when the UI changes, report coverage gaps and write release reports; a human approves before anything ships. It connects to CI/CD tools, Jira, Slack and coding assistants such as Claude Code. The docs label Atto as beta.
No agents match these filters.
“#” is the position among products of the same kind (agents, frameworks, tooling) by Gid Score, our 0–100 composite of adoption, community, attention and momentum. Use the filters and sorting above, or switch to the table view. Methodology · All AI agents.
What testing agents automate
Writing end-to-end tests is slow and maintaining them is slower: every UI change breaks selectors. Testing agents attack both. They generate tests from a description, a recorded flow or a requirements document, run them in parallel on hosted browsers and devices, and "self-heal" tests when the page changes instead of failing on a renamed button.
Questions to ask
- Do you own the tests? Some tools produce standard Playwright or Appium code you can keep; others keep tests in their own format.
- Who maintains them? Some vendors sell a managed service with QA engineers; others are self-serve.
- How is it priced? By test runs, credits, runner minutes or number of tests under management.
Coding agents can also write unit and integration tests in your repository; see AI agents for testing.
Head-to-head comparisons
Related guides
Questions about AI testing and QA agents
Can AI write end-to-end tests?
Yes. The products in this category generate end-to-end tests from natural-language descriptions or by exploring the app, and keep them working as the interface changes. Review generated tests the same way you review code: a test that always passes is worse than no test.
What is self-healing test automation?
When a page changes, for example a button is renamed or moved, a self-healing test finds the intended element again instead of failing, and records the change for review.