Virtuoso QA alternative: Plain English on top. Five selectors underneath.
Virtuoso QA (formerly SpotQA) is one of the most enterprise-serious natural-language testing platforms there is, and this page treats it that way. The real split is not the English — it is what gets stored under it, where the browser lives, and who decides what a failure means.
Last reviewed · See a real graded run
What Virtuoso QA genuinely gets right
Evidence a compliance team can live in
Per step: a screenshot of the last pass next to the failing run, the rendered DOM, the full network log with 4xx/5xx badges, console output, a 90-day run history. Add SOC 2 Type 1 and Type 2, SSO/SAML, and requirements-to-test traceability — a regulated buyer's checklist, mostly checked.
A journey model with real depth
Project → Goal → Journey → Checkpoint → Step, with shareable library checkpoints, versioning, environments, and API calls made inside UI journeys. Runs fan out across Chrome, Edge, Safari and Firefox on their cloud grid — zero infrastructure on your side.
Authoring that respects manual testers
Live Authoring validates each English step against a real headless browser as you write it; StepIQ corrects typos and suggests next steps; a JavaScript Extensions escape hatch covers the hard 10%. Their GENerator claims to convert legacy Selenium, Tosca, TestComplete, Cypress and UFT suites in — their marketing, but a real on-ramp story.
A locator, five ways
Virtuoso's docs are admirably candid about what your English becomes. A step compiles to a structured action — click, write, assert — plus a target element and an optional value. And their elements documentation spells out what a target element is: a stored stack of selectors — Hint, XPath, ID, CSS and JS path — tried in order, first match wins.
Self-healing exists because that stack exists. When the app drifts, Virtuoso keeps your Hint selectors, prunes the dead ones, and suggests new ones; high-confidence heals apply automatically (with a reject option), low-confidence ones wait for your acceptance. The oft-quoted "95%"? Their own FAQ defines it as the rate at which users accept healed steps — an acceptance rate, not an independently measured healing accuracy.
- Their artifact: English compiled to action + element steps; each element a five-type selector stack, pruned and re-suggested as the app changes.
- Our artifact: intent — a prompt and an expected outcome in versionable YAML. There is no selector layer anywhere in the file, so there is no stack to prune and nothing to heal.
- Why it matters: a healed selector is a guess that the new element is the old one — "probabilistic determinations", in their own words. An intent-driven run never guesses backwards: it re-reads the live page against the stated expectation, every run.
To their credit, a step that cannot be healed with confidence fails rather than guessing silently. The structural point stands: maintenance shrinks when the thing that decays is repaired well. It disappears when that thing is never stored at all.
A tunnel is not a door key
Virtuoso's FAQ is plain: the platform is hosted on AWS and interacts with your application using a headless browser — "no software, apps, or extensions required." For a public site, that is a genuine convenience.
For the app your team actually ships — behind a VPN, behind Okta, full of real staging data — their answer is Virtuoso Bridge: a client you run inside your network that opens a TLS reverse tunnel out to their Bridge Server, so their cloud bots can forward requests back in. It works, and its limits are documented honestly: no official macOS client, localhost unreachable, throughput bound by the client machine. A Fixed IP option covers firewall allowlisting. Either way, execution never leaves their cloud, and reusing your live logged-in session is not documented — the cloud browser signs in with steps.
- TestAutomate drives your actual Chrome. Real logged-in sessions, behind your VPN and SSO, on real staging data — because the browser doing the testing is the one that already has access. No tunnel client, no allowlist ticket, no test-only auth path.
- Unattended runs stay honest. A named Auth Catalog (email/password, Google, Okta, magic links, TOTP) handles sign-in for scheduled runs, with a live view and click-to-take-over when SSO or 2FA needs a human.
- Mid-run questions go to you. The agent can ask a human when reality is ambiguous, instead of guessing in a data center.
A reverse tunnel is security work you do so an outside browser can imitate the access your own machine already has. We think the simpler answer is the browser you already trust.
Who classifies a failure?
Virtuoso's per-step evidence trail is among the best documented in the category. The verdict grammar on top of it is thinner: a step passes or it fails, and when a test fails, you are prompted to answer why — a Regression? An Outdated test? Classifying the failure is a human's job. Auto-retry and flake quarantine are not documented; their 2025 agentic marketing says agents "classify failures, suggest repairs and rerun tests" behind approval gates — their claim, not in the docs we could verify.
- Where does a broken environment land? In a pass/fail world, a dead staging server and a product bug arrive the same color, and a person sorts them out afterwards. Our verdicts are pass / fail / blocked, graded by the system — a broken environment can never be reported as a false fail.
- Who confirms a failure? On our side, no failure is reported until a second, stronger model re-drives the journey and agrees. Nobody else does two-model failure confirmation.
- Is a retried pass labeled? Not documented on their side — retries themselves are not documented. On ours it is mandatory: a pass on retry is marked "passed on retry", never laundered into a clean green.
On the bug loop, credit first: Virtuoso's Jira app is genuinely two-way — issues link to step-level results, update in both directions, and close on final status. That beats most of the category. What is not documented is the step before filing — searching the tracker for an existing duplicate — and the two-way loop documented is Jira's (their features page also names TestRail for defect tracking, plus a Slack integration). Our loop runs the whole circle on Jira or Linear: dedupe against the live tracker and a cached inventory, file with evidence attached (always confirm-gated), re-verify the fix, transition the issue closed — plus sprint UAT that turns a board query into a durable work queue.
Four design choices, one theme
Store intent, not bindings
Prompt + expected outcome in portable YAML. No selector stack is ever recorded, so app drift changes nothing about the artifact — the run re-reads reality each time.
Test from inside the wall
Your Chrome, your session, your VPN. The access problem their Bridge tunnel and Fixed IP allowlists exist to solve simply does not occur.
Grade, don't just color
Pass, fail, blocked, advisory flags, labeled retry-passes — the system classifies, and a second model must confirm every failure before you hear about it.
Charge nothing on the meter
Bring your own key and model spend passes through at provider list price with a live dollar meter and a hard ceiling. No key? Managed credits at provider price + 15%, both numbers shown.
Side by side
| Virtuoso QA | TestAutomate | |
|---|---|---|
| What you write | Plain English, live-validated as you type (Live Authoring), with StepIQ suggestions | Plain English — typed or voice-dictated |
| What is stored & replayed | English compiled to action + element steps; each element keeps a selector stack — Hint, XPath, ID, CSS, JS path — tried in order, first match wins | Intent: prompt + expected outcome in versionable YAML — no selector layer exists in the artifact |
| Self-healing | Hint-guided repair of the selector stack; high-confidence heals auto-apply; "95%" = user acceptance of healed steps (their FAQ) | Nothing to heal — no stored binding; renamed-but-equivalent UI passes with an advisory flag |
| Where tests execute | Their AWS cloud, headless — Chrome, Edge, Safari, Firefox; mobile-web coverage claimed on their features page, native mobile apps on their roadmap (their FAQ) | Your own Chrome; self-hosted runner or cloud browsers for unattended runs |
| VPN'd staging & private apps | Virtuoso Bridge reverse tunnel (no official macOS client, no localhost) or Fixed IP allowlisting — execution stays in their cloud | Already inside — it is your browser, your session, your network |
| Logged-in session state | Not documented — the cloud browser signs in with scripted steps | Real logged-in sessions, plus a named Auth Catalog for unattended sign-in |
| Verdict grammar | Pass or fail per step; you are prompted to classify why a test failed; auto-retry/flake quarantine not documented | Pass / fail / blocked graded by the system; advisory flags; "passed on retry" labels; a stronger model confirms every failure |
| Evidence per verdict | Excellent: last-pass vs failing screenshots, rendered DOM, network log with 4xx/5xx badges, console, 90-day history | Scrubbable replay + itemized verdict per expectation, share-by-link, PDF, evidence attached to filed bugs |
| Bug-tracker loop | Jira two-way sync — issues linked to step results, close on final status; TestRail also named for defect tracking on their features page; dedupe-before-filing not documented | Dedupe against the live tracker first, confirm-gated filing with evidence, close-fixed after re-verification — Jira or Linear |
| Pricing | No public numbers — quote-based Consumption or Capacity models; standard 12-month term (their pricing page, as of Sep 1, 2026) | BYOK at provider cost — never marked up — live meters, hard budget ceilings; managed keys at provider price + 15%, shown |
| Test export | Proprietary JSON backup of a goal (8MB import cap); code export not documented — GENerator converts suites into Virtuoso, not out | Portable human-readable YAML, one-click full-suite export |
| Self-hosting | Not documented — hosted on AWS (their FAQ); SOC 2 Type 1+2 certified | Entire stack self-hosts on Node + Postgres — self-host and BYOK together |
One row deserves a footnote: Virtuoso publishes its pricing structure honestly — a Consumption model (authoring users plus execution volume) and a Capacity model (parallel execution lanes) — but no numbers anywhere, and the standard contract runs twelve months. The checkout is a custom-quote form. Ours is arithmetic you can check against your provider's price list.
TestAutomate vs Virtuoso QA, asked directly
Isn't Virtuoso QA selector-free too?
No — and their docs don't pretend it is. Your English is compiled to structured steps, and each target element stores a stack of selectors — Hint, XPath, ID, CSS and JS path — tried in order at runtime, with a "smarter analysis" fallback. Self-healing manages that stack when the app drifts. TestAutomate stores only the prompt and the expected outcome; there is no stack to manage, so there is nothing to heal.
What about their 95% self-healing number?
Quote it precisely: their FAQ says the healing has "been proven incredibly accurate, as the self-healed steps are accepted by our users 95% of the time." So the number itself is an acceptance rate — how often humans click yes on a proposed heal — offered as the evidence of accuracy. We publish no healing number at all, because our artifact has no bindings to heal.
Can Virtuoso QA test apps behind a VPN?
Yes — via Virtuoso Bridge, a client you run inside your network that opens a TLS reverse tunnel so their cloud bots can reach in, or via Fixed IP allowlisting. Their docs list the limits: no official macOS Bridge client, localhost unreachable, throughput bound by the client machine. The difference is ergonomic: that is infrastructure you stand up so an outside browser can get in. TestAutomate runs in your own Chrome, which is already in.
Which one costs less?
We genuinely cannot tell you, because Virtuoso publishes no prices — as of September 1, 2026 their pricing page describes a Consumption model and a Capacity model, a standard 12-month term, and a custom-quote CTA. TestAutomate passes model spend through at your provider's list price on your own key, with a live meter and a hard ceiling; managed credits show provider price + 15% explicitly.
Can I take my tests with me?
Virtuoso's documented export is a proprietary JSON backup of a goal and its entities, with an 8MB import cap — fine for backup, but no export to Selenium, Playwright or any code format is documented, and their GENerator migration tooling points the other way: into Virtuoso. TestAutomate tests are human-readable YAML with one-click full-suite export, and the whole stack (Node + Postgres) self-hosts. Leaving us is a copy, not a rewrite.
Skip the tunnel entirely
Point TestAutomate at the staging environment behind your VPN, in your own logged-in Chrome, and watch a graded verdict arrive with a replay attached — at your provider's list price.