Writes tests from plain language
Paste a test case the way your team already writes it. The AI step writer drafts the steps, reuses the fields your tests already know, and shows the sentence each step came from.
Learn more →AI test automation for web applications
AI that writes, heals and explains your tests.
Describe a test the way your team already writes it. AutoTestX’s AI drafts the steps, heals them when the application changes, and explains every failure, while each run stays deterministic and keeps a screenshot of every step.
For any web application. Running IBM Maximo? See the Maximo edition →
Test case, as your team wrote it
- 1Sign in to the customer portal as a support agent
- 2Open a new service request
- 3Choose the category Facilities
- 4Describe the problem: Heating not working on floor 3
- 5Submit the request
- 6Check the message says the request was created
✦ Draft steps
- Sign in as
{{secret.AGENT_USER}}from line 1 - Click
New requestfrom line 2 - Select Category
Facilitiesfrom line 3 - Type Description
Heating not working on floor 3from line 4 - Click
Submitfrom line 5 - Expect message
Request SR-… createdassumed · check
✦ 6 steps drafted1 assumption to reviewReview and save
Run #2187 · Submit a service requestStaging
- 01Open
Customer portal✓ - 02Sign in as
{{secret.AGENT_USER}}✓ - 03Click
New requestHealed: #btn_new_7f2 → button “Create request” · 96% match · waiting for your approval✦ - 04Select Category
Facilities✓ - 05Type Description
row 3 of requests.xlsx✓ - 06Click
Submit✓ - 07Expect message
Request SR-1042 created✓
7 passed✦ 1 healed · approve fix7 screenshots
› Run the Checkout tests on Staging and tell me if anything broke.
autotestx · run_suite Checkout on Staging · 12 tests
11 passed · 1 failed · 2 min 41 s
autotestx · explain_failure Apply discount code
Apply discount code failed at step 4: it expected “Discount applied”, but the page said “Code expired”.
The button and the form worked, so this is the test data: the promo code has expired. It passed until 29 Sep. Renew it in the staging seed data.
Failure class · Test dataEvidence · step 4 screenshot
Illustrations of the AI step writer, a healed run, and an AI assistant using the AutoTestX MCP server.
Works with
- Any browser application
- Claude Code
- Cursor
- VS Code
- GitHub Actions
- GitLab CI
- Jenkins
- Azure DevOps
- Excel and CSV
- REST APIs
- Windows and macOS
Watch the intro (2:09)
See what AutoTestX does.
A test case written, run, explained and healed, in the real product on sample data.
Recorded in our demo environment with real AI calls. Northwind Supply is a practice application; the data is sample data and the narration is a synthetic voice. See the full product demo →
The AI in AutoTestX
Four AI capabilities. One platform.
Each one takes repetitive work off your team, and each one hands the decision to a person.
Heals tests when the app changes
A renamed button or a moved field no longer breaks the run. AutoTestX finds what the step meant, keeps going where that is safe, and files the fix, with its evidence, for approval.
Learn more →Explains every failure
Each failure is classed (the application, the environment, the test data or the test), and AI diagnosis explains the likely cause in plain language, with the evidence it rests on.
Learn more →Answers your AI assistant
Claude Code, Cursor or VS Code run your suites through the AutoTestX MCP server and tell you what broke, without leaving the editor. In an editor, they can fix the code and run the suite again.
Learn more →AI you can audit
AI suggests. People decide. Runs never guess.
Every step of a run executes exactly as written, so the same test on the same data gives the same result. Here is what each AI feature sees, never sees, and who makes the call.
AI AI step writer
- Sees
- Your test case text and the names of fields your tests use
- Never sees
- Secret values
- Who decides
- A person reviews the draft before it is saved
AI AI healing
- Sees
- The step, and candidate elements’ roles, labels and ids
- Never sees
- The page, screenshots, typed values, secrets
- Who decides
- Filed as a suggestion; never acted on during a run
AI AI diagnosis
- Sees
- The step, the error and failed requests
- Never sees
- Screenshots, recordings, secret values
- Who decides
- Your team decides what to fix
AI AI assistants (MCP)
- Sees
- What the AutoTestX API returns to your token
- Never sees
- Anything your token can’t reach
- Who decides
- You ask; the run is a normal run
From test case to release
AI at every stage of testing.
One platform from a test case on paper to the release decision.
- 01 · BuildAI step writer
Turn a written test case into an automated test
Record the flow in your own browser, or paste the case and let the AI draft the steps. Data from Excel, CSV or an API: one run per row.
Learn more →
- 02 · RunAI assistants (MCP)
Run on every release, without anyone clicking
Suites run from the app, on a schedule, after each deployment from your CI pipeline, or when your AI assistant asks.
Learn more →
- 03 · InvestigateAI diagnosis
Know why it failed, not just that it failed
The failed step, expected versus actual, the screen at that moment, video and trace, a failure class, and an AI explanation.
Learn more →
- 04 · MaintainAI healing
Keep tests current as the application changes
Self-healing proposes fixes with a confidence score; rules can apply safe ones; every decision is audited and reversible.
Learn more →
- 05 · Release
Make the release call with the evidence in front of you
A dashboard, run history, and a release report that separates new regressions from known failures.
Learn more →
Editions
One AI platform. A dedicated edition for Maximo.
For any web application
Customer portals, ERP and CRM screens, internal tools, SaaS: anything your testers reach in a browser.
- AI step writer, AI healing, AI diagnosis, MCP
- Recorder and plain-language step editor
- Data-driven, API and custom code steps
- CI, schedules and evidence reports
The Maximo edition
The same AI platform, with what IBM Maximo and MAS teams need built in, and pilots run by people who know Maximo.
- Lookups, tabs, frames and table rows by business key
- BMXAA messages turned into checks
- Failures classed as Maximo business rules
- Fix packs, upgrades and the move to MAS
Hosted in Canada (AWS) · Agent with no admin rights · Secrets redacted from logs and reports · Five roles per project · IT pre-check →
FAQ
Questions teams ask first.
Everything the product does today, and doesn’t yet, is in the capability list.
What does the AI actually do?
Four things: it drafts test steps from plain-language cases, picks the moved element when healing can’t prove one, explains why a test failed, and lets your AI assistant run suites through MCP. Each is a suggestion a person reviews.
We can’t have AI making decisions in our tests.
It doesn’t. Runs are deterministic: each step does exactly what is written, and no run acts on an AI answer. An organization admin can turn AI off altogether.
Which AI does it use?
Anthropic’s Claude, on the AutoTestX key or your organization’s own Anthropic key. Screenshots, recordings and secret values are never sent to it.
Which applications can it test?
Web applications your testers reach in Chrome or Edge, including ones behind your VPN for recording. There is a dedicated edition for IBM Maximo. Native mobile and desktop applications are not supported.
Do our testers need to code?
No. Tests are steps in plain language, drafted by AI or recorded. Custom code and API steps are there when a developer wants them.
What does it need from our IT team?
A per-user agent for recording, with no admin rights, and for unattended runs a test environment the AutoTestX server can reach. The IT pre-check covers it.
Request a session
Bring one workflow. See the AI test it.
In 30 minutes: a live demo of AI drafting, healing and explaining tests on a representative flow, and what your team would need for a pilot.