Tier 1 — Aurora Desktop App

Your whole test team,
generating scripts — not just the lead.

Aurora is the AI-powered test authoring layer of the Aurora Platform — built so every tester on the team can generate and maintain automation, not only the automation lead or a single senior engineer. It ingests your existing test documentation, builds structured test packs with automatic locator healing, and pushes everything into Continuum — ready to run on every change.

See early adopter pricing Book a demo

Early adopter pricing — 20% off Starter & Pro, ends 30 Sept 2026 New: Team licence — up to 6 testers on one plan

Windows macOS Linux

The problem

Most scripting time isn't testing. It's transcription.

A senior test engineer in a typical enterprise environment spends 60–70% of their time writing and maintaining test scripts. Not exploring edge cases. Not designing test strategy. Writing code that converts a documented test case into something executable — then fixing it every time a locator breaks.

That's the problem Aurora solves. Not by generating tests from scratch, but by taking what your team has already documented and turning it into structured, maintainable test packs — automatically.

60%

Of test engineer time spent on scripting and maintenance

£60k+

Fully loaded annual cost of a senior test engineer

Hours

To build a test pack that previously took weeks

Auto

Locator healing when the UI changes under you

Benefits

Built for developers who ship fast

Whether you're a developer adding coverage as you build, or an SDET owning the test suite — Aurora gets out of the way and lets you move quickly.

From zero to running test in minutes Describe what you want tested. Aurora probes the page, plans the scenarios, and generates a working Playwright spec — no boilerplate, no selector hunting, no framework setup.

Coverage tied to what you're shipping Test packs sit under specific features, and features sit under delivery initiatives. You always know which parts of your work are covered — and where the gaps are — without maintaining a separate spreadsheet.

Know which feature broke, not just which test When a spec fails, the failure maps back to the feature it covers. No reverse-engineering which part of the product is affected from a stack trace — the context is already there.

Bug reports that write themselves Failed runs generate AI-written bug reports with the error context already captured and linked to the feature that broke. One click to drop into your issue tracker — no manual write-up at the end of the day.

Tests that fix themselves when the UI changes The Heal stage detects broken selectors and regenerates the spec automatically. Spend your time writing code, not chasing test maintenance after every UI update.

Catch regressions before the next release Schedule test packs as synthetic monitors against staging or production. Always-on coverage between deployments — without adding another CI pipeline stage or monitoring subscription.

On-device AI pipeline The full Probe → Plan → Generate → Run → Heal pipeline runs on your machine. Point Aurora at a local model via any OpenAI-compatible endpoint and the entire pipeline runs air-gapped — no AI traffic leaves your network. Or use a cloud provider; the choice is yours and can be changed at any time.

Playwright-native output Generated specs use best-practice Playwright patterns — async/await, role-based selectors, and structured page interactions. Drop them straight into your existing test suite with no rework.

Self-healing tests When a spec fails, the Heal stage analyses the error output and produces a corrected version automatically. Fewer broken tests blocking CI, less time spent on manual selector fixes.

Synthetic monitoring built in Schedule any test pack as a synthetic monitor. Continuous pass/fail tracking with failure alerts — no separate monitoring service, no additional configuration.

Continuum integration Push results to Aurora Continuum automatically for cross-team dashboards, DORA metrics, and executive reporting. Available on Starter and Pro tiers.

SQLite on-device storage All test history, results, and configuration are stored locally in SQLite. No database server to manage, no cloud sync, no telemetry. Your data stays on your machine.

AI cost & quality visibility Every generation stage logs token usage and estimated cost against your API key. Acceptance rates and heal frequency are tracked across your test packs — giving you clear visibility into what your AI spend delivers, with no black box.

How it works

From existing documentation to executable test packs.

Aurora is a desktop application built on Tauri and Rust, calling the Anthropic API directly. It runs on your machine, inside your perimeter — no external data transfer except the AI generation call you control.

Ingest your test cases

Aurora reads your existing test documentation in whatever format your team maintains it — Word, Markdown, spreadsheet, or test management tool export. It doesn't require you to change how you work before you start.

AI-powered pack authoring

Each test case is processed through your chosen AI model, generating a structured Playwright spec with correct selectors, assertions, and execution flow. Packs are organised under Initiatives and Features, so every spec maps to a business objective from the moment it's created.

Automatic locator healing

When your UI changes and selectors break, Aurora heals them. It identifies the element by context, repairs the locator, and keeps your test pack current without engineer intervention.

Push to Continuum

Completed test packs are pushed to Continuum for remote execution with a single click. Pack variables are resolved intelligently at push time — plain values substituted inline, sensitive credentials encrypted at rest and decrypted only inside the Playwright Worker at run time. Nothing sensitive ever touches a spec file on disk.

Capabilities

What Aurora gives your engineering team

Everything needed to take test documentation to a running, self-healing test pack — without changing how your team works.

Any structured documentation format Word, Markdown, spreadsheets, or test management tool exports — Aurora ingests whatever format your team already uses.

Your AI provider — or none at all Choose Anthropic Claude, OpenAI, Azure OpenAI, or point Aurora at a local model via any OpenAI-compatible endpoint. AI calls go directly to your chosen provider with no Selectred middleware. Run fully air-gapped with a local model and no AI traffic leaves your network.

Automatic locator healing When application UI changes and selectors break, Aurora heals them automatically. No manual maintenance between releases.

Gap identification Aurora flags test coverage gaps against your documented scenarios — so you know what's missing before it matters in production.

Synthetic monitoring Schedule any test pack as a synthetic monitor. Continuous pass/fail tracking with alerts on failure — no separate monitoring service.

Direct Continuum push & run Push completed test packs to Continuum for remote execution with one click. Sensitive pack credentials are encrypted at rest inside Continuum and decrypted only in-memory inside the Playwright Worker — never written to disk in the clear. Starter and Pro tiers.

Editable AI prompts Starter and Pro: customise the system prompts at every pipeline stage to match your stack, conventions, and selector strategy.

Playwright-native output Generated specs use best-practice Playwright patterns — async/await, role-based selectors. Drop straight into your existing test suite.

AI cost & quality visibility Every generation stage logs token usage and estimated API cost against your key. Track acceptance rates and heal frequency across your test packs — so you can see exactly what your AI spend delivers and where to refine your prompts.

“Aurora isn't a co-pilot. It's the first stage of an integrated quality pipeline — turning what your team already knows into something your CI/CD can run on every change.”

— Martin Croft, Founder, Selectred

Next in the platform — Tier 2

Aurora pushes test packs into Continuum

Continuum executes your test packs and aggregates results across all frameworks. It tracks change signals, maps quality to business initiatives, and feeds DORA metrics — automatically, on every change.

Explore Continuum →

In action

See Aurora for yourself

From test generation to synthetic monitoring — everything in a single desktop app.

AI Test Generator

Aurora test packs organising tests by initiative

Test Packs by Initiative

Monitoring Overview

Pricing

Plans for every team size

No sales call required for Starter or Pro — pick a plan and check out in minutes. Got a team of testers rather than one automation lead? See Team below.

Early adopter pricing — 20% off Starter & Pro

Locked in for your first annual term. Ends 30 September 2026 — 37 days left.

Save 20%

Starter

£2,000/yr

£1,600/yr

For QA managers running multiple product areas with active monitoring.

15 features
20 initiatives
25 test packs
25 synthetic monitors
AI test generation
Edit AI prompts
Continuum integration

Get Starter

Common questions

Everything you need to know about Aurora.

Aurora supports Anthropic Claude, OpenAI, Azure OpenAI, and any OpenAI-compatible local model — Ollama, LM Studio, and similar. You provide your own API key or point Aurora at your local endpoint. No data is routed through Selectred servers. For teams with strict data residency requirements, a fully local setup means no AI traffic leaves your network at all.

Aurora runs five stages automatically: Probe (live page structure extraction), Plan (AI generates numbered test scenarios), Generate (AI writes a Playwright spec), Run (executes the spec), and Heal (if the spec fails, AI diagnoses and fixes it).

Open Aurora → Settings → System → License → paste your key → click Activate. The tier badge updates immediately. No restart required.

Both tiers include unlimited AI test generation. Starter raises all limits (15 features, 20 initiatives, 25 test packs, 25 monitors) and unlocks AI prompt editing and Continuum integration. Pro removes all caps entirely and adds priority support.

The Team plan is all of Pro, licensed for up to 6 concurrent testers instead of one machine — so anyone on the team can describe a scenario and generate a maintainable Playwright test pack, not just the person who owns automation today. Request a quote and we'll get you set up.

Aurora runs on Windows 10+, macOS 10.15 Catalina+, and major Linux distributions (Ubuntu 22.04+, Fedora 38+).

Starter and Pro licenses cover one active machine at a time. To move to a new device, visit aurora.selectred.com/account, deactivate the current machine, then activate on the new one — managed through your account portal, not from within the app. The Team plan covers up to 6 active machines at once, so multiple testers can each generate and run test packs in parallel.

Email support@selectred.com with your purchase email and we'll resend your key immediately.

Support