Agent QA

Agent-nativeSource Available

Autonomous QA agent for natural-language web and mobile application testing

Visit website GitHub

Overview

About Agent QA

This summary uses submitted and public product material. Only source-mapped claims are documentation reviewed.

Agent QA is an autonomous QA agent that authors, runs, and troubleshoots natural-language tests for web and mobile applications. It exposes CLI, MCP, and agent-skill interfaces, and uses execution memory to adapt tests when an application’s UI changes.

The current release is source-available under FSL-1.1-ALv2. That license permits copying, use, modification, and redistribution for purposes other than a competing use, and each release converts to Apache-2.0 after two years. Configured model, browser, or device providers may charge separately. Source and license terms are available in the Agent QA repository.

So agents can…

Author and run web or mobile application tests from natural-language instructions
Invoke QA workflows through MCP or reusable agent skills
Reuse execution memory when adapting tests to changed interfaces

Listing decision

Why Agent QA is listed

Agent QA uses a model-driven execution loop to observe applications, plan and perform test actions, recover from UI drift, and retain execution memory, making the agent itself central to the product.

Before you choose

Fit and limitations

Best for

Teams that want reviewable natural-language end-to-end tests across web, Android, or iOS and want coding agents to author, run, and triage those tests through CLI, dashboard, MCP, or skills.

Known limitations

The CLI and dashboard require a JavaScript runtime and configured LLM; normal visual testing needs a multimodal model. Docker is required for isolated hooks, and mobile runs also depend on the relevant Appium and platform tooling.

Not yet verified

The first-party pages describe self-healing and usage savings, but this review did not find or validate an independent accuracy, reliability, or cost benchmark. Model and device-provider charges are outside Agent QA.

Sources

Sources and verification

First-party unless marked otherwise. Each link supports only its adjacent claim.

Agent QA Quickstart — runtimes, models, and local installation
The quickstart documents local npm installation, web and mobile targets, multimodal-model requirements, CLI and dashboard operation, and optional Docker-isolated hooks.
official documentation · accessed Sep 3, 2026
Agent QA MCP documentation — agent authoring and run tools
Agent QA exposes HTTP and stdio MCP servers with tools for authoring tests, queuing runs, inspecting evidence, and classifying failures.
official documentation · accessed Sep 3, 2026
Agent QA product page — features and free local use
Vostride describes natural-language web, Android, and iOS tests, self-healing execution, file-backed memory, and free installation with model inference as the running cost.
official product page · accessed Sep 3, 2026
Agent QA license — FSL-1.1-ALv2
The current release uses FSL-1.1-ALv2, excludes competing use, and grants an Apache-2.0 future license after two years.
official license · accessed Sep 3, 2026