Back to Gap Map Product / UX / Orchestration & agents

Agent QA

Vostride
source available / Overall score: 1.7

Agent that tests web and mobile apps from tests written in natural language. It plans and carries out each action, re-observes the interface and tries another path when a step fails, and keeps memory from earlier runs to reuse proven actions. It ships a CLI, a dashboard, MCP and agent skills, and runs on the model the user brings.

Openness

2 high confidence
2.0
license
FSL-1.1-ALv2 (Functional Source License, not OSI; converts to Apache-2.0 two years after each release, and restricts competing use until then)
source
public (full TypeScript monorepo, CLI, dashboard and packages)
core features withheld
no

Agent QA is under the Functional Source License, which converts to Apache-2.0 two years after each release and forbids competing use until then, so today's code is source-available rather than open source. The whole harness, CLI and dashboard are published in one repository, and the vendor states that no feature is held back behind a paid tier.

Adoption

1 medium confidence
1.0

Adoption is measured by npm downloads of agent-qa, the product's declared distribution channel, over the trailing month. Raw download volume includes CI and mirror traffic, so this measures install volume rather than a count of teams.

Capability

2 medium confidence
2.0

Agent QA runs one agent loop for one task, testing applications, with memory carried between runs. GPT Researcher, the closest single-task agent on the map, adds a multi-agent assistant and a recursive deep-research workflow that produces cited reports, so Agent QA scores a step below it. Neither has third-party benchmark coverage.

Verified 2026-09-26