G-TESTER

Evidence-Grounded, Doctrine-Aligned Autonomous QA Service

4 followers

Evidence-Grounded, Doctrine-Aligned Autonomous QA Service

4 followers

The era of manual QA is over. Meet G-Tester — autonomous QA-as-a-Service replacing a full Senior QA team. A plug-and-play SaaS running 20 verification disciplines on any code & cloud app. Powered by GLM 5.2 for deep vulnerability checks and heavy tooling (k6, Playwright, pa11y). Emits a complete QA package & release verdict grounded in OWASP/CWE. Strict QA ethics: unrun tests skip, never pass. 🔒 100% Private SaaS: Isolated contour, 0 data leaks to external vendors.

Payment Required

Launch tags:Software Engineering•Developer Tools•Artificial Intelligence

Launch Team / Built With

Maker

📌

Hi Product Hunt! 👋 I’m Taras, Founder & Director at AlfaCan Defence Group Limited. Software Quality Assurance is a massive bottleneck: senior QA engineers take weeks per release, cost a fortune, and produce inconsistent results. On the flip side, naive "LLM-as-a-tester" tools hallucinate bugs because they judge code without running real tests. We built G-Tester to solve this. It isn't just another scanner — it is an autonomous QA-as-a-Service platform designed to replace an entire QA engineering team. G-Tester operates in 5 structured stages: 1️⃣ Evidence Collection: Runs a heavy tool fleet (k6, Playwright for Chromium/Firefox/WebKit, pa11y, HTTP probes) across 20 specialized test disciplines to gather raw, unvarnished facts. 2️⃣ Senior AI Analysis: Domain-specialized models (ITDR-Officer, S-Coder-Reviewer) filter false positives, chain findings, and analyze business logic flaws. 3️⃣ Doctrine Grounding: Two-stage RAG (Qwen3 Embedding + Reranker) grounds reasoning in house QA standards and OWASP/CWE doctrines. 4️⃣ Complete QA Package: Delivers a full Test Plan, Bug Reports, Summary, and a deterministic Release Verdict (production_ready, ready_with_caveats, not_ready). 5️⃣ Self-Learning Debrief: Distills sanitized, generalizable lessons back into doctrine so the system gets smarter with every audit. Crucially, G-Tester is built on strict Honesty: if a tool didn't run, it is marked as a SKIP — never a pass. And if the AI tier drops offline, it gracefully falls back to deterministic findings without fabricating results. Try out our live studio dashboard at https://studio.alfa-can.com (or https://guardian.alfa-can.com/st...) and let us know your thoughts and questions below!

Report

1mo ago

Maker

When we were designing G-Tester, we explicitly made sure to account for every single edge case — from zero-hallucination multi-model debates to real-time human steering in the AI Boardroom.

If you have any technical questions about the architecture, our test runners, or how G-Tester can replace manual QA in your workflow, feel free to ask right here in the comments or chat directly with our team!

We are live and will be happy to answer all your questions. Let’s talk QA!

Report

1mo ago

Reviews

Maker

📌

Report

1mo ago

Maker

When we were designing G-Tester, we explicitly made sure to account for every single edge case — from zero-hallucination multi-model debates to real-time human steering in the AI Boardroom.

We are live and will be happy to answer all your questions. Let’s talk QA!

Report

1mo ago