Gua vs Razer QA Companion-AI

Gua is a free, open-source protocol for testing and automating games made with Godot and Unity. It exposes semantic interface state, including identifiers, roles, text, hierarchy, bounds, visibility, enabled state, values, focus, and selection, so tests and AI agents can query controls and perform actions such as clicking, focusing, setting values, selecting, scrolling, and pressing keys. Tests can wait for interface conditions in NUnit, xUnit, MSTest, or C++ without fixed sleeps. Visual testing compares PNG screenshots with baselines using pixel thresholds, difference ratios, masks, and operating-system or renderer variants. Recording stores semantic actions, timing, revisions, and wait conditions for replay with host checks. Failure artifacts may include UI trees and diffs, pending requests, event history, logs, runtime and environment details, screenshots, and game output. Native MCP connects over WebSocket; WebMCP exposes tools in Godot Web Export and Unity WebGL. Gua documents a Godot 4.7 add-on and Unity 6 support. It targets Windows, Linux, macOS, API, and web environments.

Razer QA Companion-AI monitors gameplay sessions for game testing teams and turns detected issues into reviewable reports. It tracks FPS, CPU, GPU and memory use in real time, flagging breaches of thresholds defined by the team. Vision AI examines captured gameplay video and screenshots for visual glitches, physics or collision problems, interface inconsistencies, missing assets and sequencing errors. The service can draft bug tickets with descriptions, expected and actual results, and suggested reproduction steps for human review. Teams can generate functional, boundary and negative test cases from prompts or uploaded Game Design Documents. Jira is the documented ticketing integration, supporting direct submission and duplicate detection. Razer says no SDK, plugin or game-side instrumentation is required. Setup uses the web platform and a Game Data Bridge on a Windows 10 or 11 test PC, with stable internet. Pricing is on request.