Skip to main content
L3gitUI & API TEST AUTOMATION
[ PRODUCT ][ PRICING ][ DOCS ][ VERIFIED ][ REGULATED ][ COMPARE ][ ACCOUNT ] Start Free Trial

Release notes

What's new in L3git

Every release, in plain language, with how to turn each thing on. Existing installs are offered updates in-app (About → Check for updates); updates keep your key, settings and recordings.

1.1.0

2026-11-03

Scale without a Selenium Grid

Jenkins Matrix pipeline, written for you NEW

Generated projects now ship a declarative Jenkins matrix pipeline: shards × browsers, every cell on a Jenkins agent with your label, each driving the browser installed there and running only its slice of the suite. Scenarios are split by hashing their stable @l3id tag, so the same test always lands on the same shard and adding tests never reshuffles the rest. No hub, no nodes, no cloud minutes — see Selenium Grid alternative.

How: Setup → CI / Maven project → "Jenkins matrix": shards, browsers, agent label

Deterministic shards on any CI NEW

ShardSelector is a one-file Java class in your repo: k/n in, Cucumber tag expression out. The generated GitHub Actions workflow uses it as a matrix; GitLab or anything that can run N jobs can call it the same way. --balance for perfectly even counts.

How: java -cp target/test-classes com.example.support.ShardSelector 2/4

For regulated and government teams

Section 508 audit of every page NEW

One checkbox at startup. Every page your generated tests visit is audited with the bundled axe-core engine (fully offline) against the Section 508 + WCAG 2.0 A/AA rule sets. Each page's result is attached to its scenario in the Cucumber and Allure reports, and a standalone Section 508 Compliance Audit chapter is written to target/a11y/section508.html on every run. Report-only by default; a11y.audit.fail=true makes a violation fail the scenario.

How: Setup → CI / Maven project → "Section 508 audit of every page + report"

Local AI via Ollama NEW

The authoring and repair assistant can now run on an open model on your own PC or LAN. No API key, no cloud, nothing leaves the machine — the option for networks that cannot send code or page content to a hosted model. Bring-your-own Anthropic key is unchanged for everyone else.

How: Settings → Provider → "Local model via Ollama" → Test → Save (install Ollama, then ollama pull llama3.1:8b)

Drives your installed Edge — never downloads a browser

Air-gapped zips now bundle the matching Edge WebDriver, and the app accepts any bundled driver of the same Edge major, so a managed Edge update does not strand an offline install. Online builds still fetch a driver only after announcing the exact URL, and you can supply one by hand instead.

How: nothing to do — built into the 1.1 air-gapped zips

Recorder

Steps first, five tabs, one run bar

The panel opens on your step list. The toolbox is five tabs — Record, Check, Flow, API, AI — and the Test + Generate bar stays pinned at the bottom. Settings fold away behind a gear. Same controls, a quarter of the clutter.

See what the AI picked NEW

"Fix with AI" now outlines the proposed element on the page and pins a note under the step saying what it found: tag, visible text, page region (header / nav / main / footer / form / dialog) and the HTML. If it grabbed the footer link instead of the nav link, you know before you Resume.

Simpler startup window

Essentials only: URL, name, style, browser. Driver, tags and output options live under "More options"; reports, Jenkinsfile, the log4j ban, the 508 audit and healing-DB credentials under "CI / Maven project". DB fields are disabled when you generate Plain output.

Generated projects

Stable scenario ids

Every generated Scenario now carries an @l3id_<hash> tag, minted once when you author it and persisted in the recording, so it survives renames and regeneration. This is the identity run-history and flake analysis will key on. Read it from cucumber.json or Allure results — tags are not in JUnit XML test names.

1.0.0

2026-07-19
Download 1.1 — free 14-day trial Regulated & air-gapped