← All projects

agent-qa

The self-improving Agentic QA harness with Memory

Dev Toolsqa-testingai-agentsnatural-language-testingself-healingtest-automationllmmobile-testing
agent-qa screenshot

About

agent-qa is an AI-powered QA testing harness that lets developers write tests in natural language for web and mobile applications. It builds execution memory from past runs to adapt to UI changes, self-heal failing actions, and reduce redundant work via caching. The tool integrates with popular LLMs, exposes primitives via MCP, and keeps tests as version-controlled code for team collaboration.

Problem

Writing and maintaining UI tests is brittle and time-consuming as interfaces change and tests break unexpectedly.

For

software developers and QA engineers building web and mobile applications

How it works

Developers write tests in natural language; an AI agent executes them, builds memory from each run, self-heals failures by re-observing the UI, and caches validated plans to speed up future runs.

Business model

unknown

Status

launched

Company

vostride

Similar projects