Lucius AI
Open navigation
NewLucius AI — first release

Meet your self-exploring
testing agent.

Describe it, and Lucius takes it from there — exploring your app, building the test, and running it live. No steps required.

Start testing freeTalk to us
session active
Loaded the page. Exploring it now to map the checkout flow.
Explored the flow — 3 test cases generated.
Scenario-01
Add item to cart
Pay
Confirm order
Live sandbox · connecting…
YOUR APP
Generated scenarios3

Scenario-01

Successful checkout with valid payment

Scenario-02

Checkout blocked with an empty cart

Scenario-03

Price updates when quantity changes

Create 3 test cases
Order confirmed
checkout.spec.ts
test('checkout', async ({ page }) => {
  await page.goto('/cart');
  await page.getByRole('button', { name: 'Checkout' }).click();
  await expect(page.getByText('Order confirmed')).toBeVisible();
});

Efficiency

9×

leaner output than a general LLM.

Same tests. A fraction of the tokens.

Leaner output
9×
Fewer tokens
~20%
Vendor lock-in
0

Measured against a general LLM using official Playwright.

The problem

01

Slow to write

Writing end-to-end tests is slow. Most teams either skip them or wait on QA.

02

Worse to keep

Selectors break, suites go flaky, and maintenance eats the time you saved.

03

QA bottleneck

Coverage becomes the constraint. Teams drown in tests — or ship without them.

Lucius removes all three.

All the primitives to ship tested software.

Focused building blocks for describing, generating, running, and owning real code.

01 · Intent

Intent → test

Describe the scenario once, in plain English. Lucius explores your app and turns it into real code you can run.

How it works

One pipeline. Every step visible.

Lucius isn't one model guessing at a script. It's a focused pipeline — each stage does one job, and you can see the output of every one.

01

Explore

Lucius loads your app and maps the screens and flows involved in what you described.

02

Plan

Turns intent into a concrete scenario — what to check, and in what order.

03

Code

Generates real, editable code for every step. Not a black box.

04

Verify

Checks each step as it goes, so failures are caught before you are.

05

Run

Executes the suite live and reports pass/fail per step, in real time.

The architecture

AI shouldn't own your tests.

Your automation should stay real code — even when AI creates it.

Bolted-on AI testing

AI
↓
Proprietary test format
↓
Proprietary runtime
↓
Platform

✕ Locked to the platform

Lucius

AI
↓
Explores your app
↓
Real Playwright code
↓
Your Git · Your CI · Your cloud

✓ Your tests. Your infrastructure.

Ownership

You own everything.
A tool, not a trap.

Other AI testing tools hand you a black box. When it breaks, you're stuck. Lucius generates standard, real code you can read, edit, and take with you. Cancel tomorrow and your tests still run.

checkout.spec.tsRun
import { test, expect } from '@playwright/test';

test('checkout', async ({ page }) => {await test.step('Navigate to /cart', async () => { ... });await test.step('Click 'Checkout'', async () => { ... });await test.step('Assert 'Order confirmed'', async () => { ... });
});
Console
✓Navigated to /cart
✓Clicked 'Checkout'
✓Order confirmed — assertion passed
VNC SessionConnecting…
Connecting to VNC…
YOUR APP

Security that protects your code and network. Private by default — not as an afterthought.

Privacy

Your code never trains our models by default.

Opt-in, not opt-out. What you run in Lucius stays yours unless you explicitly choose otherwise.

Default stance

Opt-in training only

Tunnel

Reach private apps without exposing them.

Tunnel connects out — nothing is exposed publicly. Test staging and prod behind the firewall.

Network

Outbound-only tunnel

Ownership

Exportable code you can take with you.

Standard, real code in Git means canceling tomorrow does not trap your suite in a black box.

Portability

Export to Git anytime

Pricing

For getting started

Free

Testing — both modes & import. GitHub and runners included.

Auto + Manual

Testing modes

GitHub

Integration

Runners

Included

Bulk Import

Test cases

Export to Git anytime
Start testing free →

For teams and organizations

Enterprise

Everything in Free, plus:

  • Custom usage limits
  • Teams & multi-organization capabilities
  • SSO & advanced security features
  • Tunnel & Browser Profiles capabilities
  • Dedicated customer support
Everything in Free, included
Talk to us

FAQ

Yes — standard, real code, exportable to Git.

Stop writing tests. Start shipping.

Describe a scenario, get real code, and ship with confidence.

Start testing free