agent-qa agent-qa
agent-qa is a self-improving, agentic QA test harness that allows users to write end-to-end tests for web and mobile applications using plain-English YAML files. It uses an LLM-backed agent to execute tests by reading UI labels and screen state, and it builds file-backed execution memory to adapt to product changes over time.
Recent versions (up to v0.1.21) have focused on refining the memory curator and expanding support for integrated testing workflows.
As of
- Natural language test authoring in YAML
- Self-healing test execution for UI drift
- Accumulated execution memory from past runs
- Sandboxed hook support via Docker
- Support for BYO LLM (OpenAI, Anthropic, Gemini, local)
Teams struggling with high maintenance overhead from brittle, selector-based test suites that want to automate linear user journeys across web and mobile.
Complex test scenarios requiring heavy conditional branching or programmatic, non-linear logic that exceeds the capability of a structured, natural-language YAML format.
The software is open-source and free to use without seat fees or vendor-hosted lock-in; users provide their own LLM credentials and infrastructure.
An excellent, highly transparent tool for teams looking to reduce maintenance 'selector rot' by shifting to adaptive, memory-backed testing that learns from experience.
Is agent-qa free?
Yes - agent-qa is Open Source. The software is open-source and free to use without seat fees or vendor-hosted lock-in; users provide their own LLM credentials and infrastructure.
What is agent-qa best for?
Teams struggling with high maintenance overhead from brittle, selector-based test suites that want to automate linear user journeys across web and mobile.
Who makes agent-qa?
agent-qa is developed by Vostride. It is listed in the AI Testing & Quality Assurance category on ai.dosa.dev.
Favorite this tool to revisit it later, or Zap it to contribute to the public vote count.
Content on this page is AI-generated. Please verify details with the vendor's website for accuracy.