Agent QA is an open-source, self-improving quality assurance agent that empowers software teams to write and execute tests in plain language across web, Android, and iOS applications. Powered by an integrated memory system known as memidex, it records facts, procedures, and run history as reviewable repository files, allowing the agent to learn from previous executions and adapt to interface changes. With dedicated visual reporting dashboards, detailed screenshot evidence for failed assertions, and native Model Context Protocol (MCP) support, Agent QA bridges human QA workflows and autonomous coding agents to catch regressions before production deployments.
주요 기능
01819 GitHub stars
02Visual failure inspection with synchronized step logs and browser screenshot evidence
03Self-improving feedback loop that retires stale knowledge and adapts to UI shifts
04Plain-language test journey definition across web, Android, and iOS environments
05Persistent memory engine (memidex) that stores validated procedures as reviewable files in your repository
06Model Context Protocol (MCP) server support for direct integration with AI coding assistants
사용 사례
01Debugging checkout flows and frontend regressions with visual, step-by-step screenshot evidence
02Running automated cross-platform regression tests written in natural language
03Empowering AI coding assistants to independently verify and test user journeys via MCP