TestPlanIt v0.21.0 adds a third source for AI-powered test case generation: any publicly accessible URL. Point the wizard at a web page, optionally crawl linked pages, and watch test cases stream in as the AI analyzes each page.
TestPlanIt v0.20.0 brings three features that make session-based and exploratory testing faster to set up and easier to share.
TestPlanIt v0.19.0 adds support for self-hosted Git servers — connecting your own Gitea, Forgejo, or Gogs instance as a code repository for AI-powered QuickScript generation.
TestPlanIt v0.18.0 ships Duplicate Detection — two complementary features that help you keep your test suite lean as it grows.
TestPlanIt v0.17.0 ships Copy/Move — select test cases or an entire folder, pick a target project, and transfer them directly. No export file, no import wizard, no field mapping. The cases arrive in the target project with their steps, custom fields, tags, issue links, and attachments intact.
Tags are one of those features that looks great in a demo and gets ignored in practice. The concept is solid: label your test cases, test runs, and sessions by area, type, or risk level so you can filter, group, and report on them meaningfully. The reality is that tagging is a manual chore, it's inconsistent across team members, and it's the first thing to slip when you're under pressure to get tests written and runs scheduled. So most repositories end up with either no tags or a sprawl of half-applied, differently-spelled labels that don't actually help anyone filter anything.
TestPlanIt v0.16.0 introduces Auto Tag — AI-powered tag suggestions that analyze your test cases, test runs, and sessions and propose relevant tags in bulk. Review the suggestions, accept what fits, and move on. Your existing tags are reused when they match; new ones are created when they don't.
Last week I wrote about why AI isn't killing manual QA — it's closing the gap between knowing what to test and expressing it as runnable code. That was the argument. Today we're shipping the solution.
TestPlanIt v0.15.0 introduces QuickScript, a new feature that converts your manual test cases into automation scripts for Playwright, Cypress, Selenium, pytest, Jest, and any other framework you use. Select cases from the Repository, pick a template, and export. QuickScript can generate complete, context-aware test files that reference your actual page objects, helpers, and utilities.
QA engineers are not test case writers. They're not script generators. They're the people in the room asking "what could go wrong?" before a feature ships — risk assessors, user advocates, and often the only ones who've read the requirements closely enough to notice the gap between what was asked for and what was built. That judgment doesn't live in a test case or a Playwright script. It lives in the person.
This is why the recurring prediction that manual QA is dying has always missed the point. The role isn't defined by its artifacts. It's defined by that instinct — the domain knowledge, the mental model of how real users behave, the memory of what broke last time. Those things are genuinely hard to replace.
The real threat to QA has never been automation. It's been stagnation — testers who stop growing, teams that treat QA as a checkbox, organizations that mistake having a test suite for having a quality culture. That's what erodes the role. Not better tooling.
A few weeks ago, I wrote about bringing your own LLM — choosing your model, using your API key, keeping your data on your infrastructure. But choosing your model was only half the equation. You still couldn't easily tune the prompts that shaped the output.
With v0.14.0, TestPlanIt ships with a full prompt configuration system — every AI prompt visible, editable, and customizable per project, directly in the admin UI.
Every test management vendor is racing to add AI-powered test case generation. And to their credit, most of them aren't charging extra for it — they're bundling it into their existing per-seat pricing. TestRail includes AI in both their Professional ($37/user/month) and Enterprise ($74/user/month) tiers. PractiTest bundles SmartFox AI into their $54/user/month Team plan. BrowserStack includes it in their paid tiers.
So what's the problem?
You're paying for AI whether you use it or not. You don't get to choose the model. You can't see or modify the prompts. You have no control over where your requirements data goes when it's processed. And in some cases — like Qase's credit-based AIDEN system at $0.40 per extra credit — you hit real usage caps when you need it most.
There's a different approach. Plug in your own LLM and API key, choose the model you trust, and keep your data exactly where you want it. Here's why that matters.