Download L3git
A native Windows desktop app. Every edition runs fully functional for 14 days — no key, no card. Activate a license anytime to keep going.
L3git UI
$49/user/moPoint-and-click browser recorder → owned Cucumber + Selenium.
Download installer Windows · .exe installer · ~16 MB Air-gapped portable .zipL3git Suite
$169/user/moBoth tools, plus the UI + API cookie-bridge test only the Suite can build.
Download installer Windows · .exe installer · ~16 MB Air-gapped portable .zipL3git API
$129/user/moBrowser-free API testing → owned REST Assured (functional + load + security).
Download installer Windows · .exe installer · ~16 MB Air-gapped portable .zipVersion 1.0.0 · digitally signed · not sure which edition? Compare editions
// L3GIT UI · API · SUITE — WHICH FORMAT?
Installer RECOMMENDED
The normal choice. Installs per-user (no admin needed), adds a Start-menu shortcut, and updates itself in-app. Pick this unless your network forbids installers.
Air-gapped portable .zip
For locked-down / regulated / offline machines. Unzip and run — it never contacts the internet (no update checks, no telemetry). Ideal for finance, healthcare, and government.
System requirements (UI / API / Suite)
- ✓ Windows 10 or 11 (64-bit)
- ✓ PowerShell — already ships with Windows
- ✓ No Python / Node / extra runtime to author tests
- • To run generated tests: JDK 17+ & Maven
After you download (UI / API / Suite)
- 1. Run the installer (or unzip and launch L3git.exe).
- 2. Start building tests — full features for 14 days, no key.
- 3. Bought a license? Paste your key on the Activation tab. Your account →
L3git is digitally signed. If Windows SmartScreen still shows a prompt on a brand-new release, click More info → Run anyway — the warning fades as the signature builds reputation.
// A DIFFERENT PRODUCT — L3GIT MODEL-EVAL (FREE BETA)
L3git Model-Eval — promotion gate for AI models · learn more
Not a desktop app and not part of the editions above: a command-line tool for AI/ML teams. Run your current model and a candidate through the same seeded scenarios, compare behavior per cluster with confidence intervals, and get a PASS/BLOCK verdict your CI enforces — it catches the model that’s better on average and broken on the slice that matters. Local-first: your models and data never leave your machines.
Its own requirements: Python 3.11+ · Windows / macOS / Linux · no installer — unzip and run l3git-model.bat init · zero dependencies, zero network calls · no trial clock — free while in beta
Classified or air-gapped environment? The engine runs fully offline and ships with an assurance document for security review — request a pilot.
ships with example gates + assurance doc