Track your testing the way
Git tracks your code.

Your tests live with your code, committed to the same repo. Every AI agent now sees the code and the tests that prove it, session after session. Testing visibility that compounds across agentic runs.

Follow the pipeline ↓

01

Proper test cases, any source

Point GTMS at a requirement and an AI adapter writes full, human-readable test cases with real intent, not the throwaway one-liners AI usually emits. GTMS owns the deterministic part: unique IDs, validation, and a traceable link back to the ticket, committed to your repo beside the code it tests.

Bring it as a file or fetch it live from your tracker. Jira today; PDF, Azure DevOps, or your own adapter next.

gtms — bash

02

Run it with AI, no script

Hand the test case to an AI agent. It drives a real browser (Chrome CDP + Playwright MCP), exercises the feature, and records a structured pass, fail, or skip. Or run the test yourself and record the result. Agent or human, same path.

No automation written yet. You get coverage today, on the tests you haven’t automated yet. Exactly where manual QA lives.

gtms — bash

03

Lock it down when it matters

When a test needs a deterministic script, graduate it. An AI agent writes it from the same intent, in whatever framework you run (Playwright here); GTMS wires it to the test case and keeps the lineage. Same test case, same history.

GTMS and the agent delivered a passing script fast. Now your coverage is auditable, reproducible regression.

gtms — bash

04

Tests pass ≠ done

A passing test can still be the wrong test. GTMS locks the human-readable test case up front and ties every script back to it. Then an independent AI reviewer (a tuned prompt) checks each script actually honours what the case asked. GTMS owns the deterministic linkage; the reviewer owns the judgement, and that split is the point.

It catches the drift CI happily went green on. Passing isn’t the goal. Honouring test case intent is.

intent-review · AI reviewer

05

See everything at a glance

One command, the whole picture. Every test case across every feature and bug-fix, with its state: passing, failing, primed, or a gap. gtms status shows the lot; gtms gaps shows what’s untested or failing.

No spreadsheet, no dashboard to maintain. It reads straight from your repo. Results live in git, so they aggregate across every AI session and travel with your code.

gtms — bash

Want a walkthrough?

Join our weekly demo and see GTMS in action live.

Training and Consulting from TestManagement.com

Test Management with Git + AI

About the "Git Test Management System"

Why the age of Traditional Test Managment Tools is over

Implementing the GTMS System

The GTMS Accelerator and GTMS Training packages are designed to help your team implement Git-based test management, migrate existing test assets, configure AI-readable specifications, and establish the processes needed for sustainable adoption across your organization.