Most test teams don't need more ideas about testing. They need time: time to write the tests the release is waiting for, to fix the ones that broke overnight, and to work out why the build is red before standup.
AI testing tools promise to give that time back, but many of them ask you to move first. Record in their format, run in their cloud, rebuild your suite on their platform. And if you ever leave, your tests stay behind.
Testver takes the opposite approach. It is an AI test engineer that works on the test framework and the repository you already have. It writes, records, runs, repairs and manages your web, API and mobile tests, and every test it touches stays plain framework code in your own repo.
Your tests stay yours
Testver sits on top of the tools your team already knows: Playwright, Selenium, Cypress, Appium, WebdriverIO, Cucumber, Robot Framework, pytest and more, in TypeScript, JavaScript, Python, Java and C#. Your tests keep running on the framework's own runner, exactly as they do today.
That has a simple consequence: there is no lock-in. Stop using Testver tomorrow and your tests still work, because they were never anything other than your tests.
Testver itself is free and self-hosted. It runs on your laptop or on your own server, and it keeps its data in a .testver folder inside your project. There is no telemetry. The only thing that leaves your machine is the AI requests you choose to make.
What Testver does
An AI Assistant that knows your project
Ask in plain English and the assistant does the work: writing a test, debugging a failure, refactoring a flaky test, calling an endpoint or building a dashboard. It reads your page objects, helpers and conventions first, so the tests it writes look like your team wrote them.
You stay in control. In Manual mode, the default, the assistant asks before it writes a file, runs a command or drives a browser. And when you only want answers, Ask mode explains how your app and your tests work without changing anything.
Record a test by using your app
With the AI Record Agent, you click through your web app (or tap through your mobile app) and Testver captures each step with resilient locators. The result is a clean test in your framework's style that reuses the page objects you already have. Need a check, a custom action or some data pulled from the page? Describe it, and Testver adds the step.
Turn requirements into test cases
Point Testver at a prompt, a PDF, a Word document, an OpenAPI spec or a Jira issue, and it drafts positive, negative, edge and boundary cases, prioritized and linked back to the requirement they came from.
Understand failures, and fix them faster
When tests fail, failure analysis groups them by root cause and explains each one in plain English. One click hands the fix to the AI Assistant; another files a defect.
When a test breaks because the UI changed, Auto-Heal reads the page the test actually saw, proposes a new locator and re-runs the test to prove the fix works before it keeps it. You choose whether to review each fix first or let verified fixes apply on their own.
Run anywhere, and see everything
Run a single test, a suite or several frameworks at once on your own machine, on Testver Grid (cloud Chromium, Firefox and WebKit with no vendor account and no keys), or on BrowserStack, LambdaTest and Sauce Labs. Every run comes back with screenshots, video, traces and console output for each step, plus the full history of every test.
For CI, your pipeline makes one call and the tests run on your Testver server, where browsers and drivers are already set up. The build waits for the verdict and the result is posted back to the commit. Templates are ready for GitHub Actions, GitLab CI, Jenkins, Azure Pipelines, Bitbucket and CircleCI.
Manage your testing in one place
Test cases, plans, cycles and executions; defects on a board with Jira built in; dashboards and reports you can email on a schedule; and Git with side-by-side conflict resolution. It is one app instead of five tools, and it can import your existing cases from TestRail, Zephyr, Xray, qTest and PractiTest.
AI your way: cloud or local
Testver gives you two ways to power its AI, and you can switch at any time:
Testver cloud AI. Frontier models chosen and tuned by Testver. One connection powers every AI feature, with no provider accounts and no API keys to manage. You prepay for credits and nothing ever auto-renews. The free Starter plan gives you credits to try every AI feature, with no card required.
A local model. Run any model with Ollama or LM Studio. It is completely free, needs no key and works offline.
Get started in a minute
You need Node.js 20 or newer (22 is recommended) on Windows, macOS or Linux. Then, from your test project's folder:
npm install -g @testver/testver
cd your-test-project
testver serveOpen http://localhost:3700. A short setup creates your admin account and asks whether the workspace is for Web or Mobile testing. Your existing tests appear straight away. There is nothing to migrate and no database to set up.
From there, three steps take you to your first AI-written test:
Connect AI. Create a free account at testver.com, activate the Starter plan, and connect it in the app under Testver → Connections. Or start Ollama or LM Studio and use a local model instead.
Record a flow. Open the AI Record Agent, click through something your users do every day, such as signing in or checking out, and review the test it writes.
Run it. Run the new test next to your existing suite and look through the results: every step, with screenshots and a trace.
The installation guide walks through each step in more detail.
Running Testver for a whole team
When the whole team wants in, you have two options. Install Testver on one always-on server behind an HTTPS proxy, so schedules and CI runs have somewhere to land. Or let us host it: Testver Cloud gives your team its own always-on Testver workspace at its own address, built, backed up and repaired for you. Cloud is an add-on for Team plans, and Grid is an add-on for any paid plan.