Documentation

Ask anything about testing strategy, coverage, or risk and get a grounded answer with sources you can check, not generic guesswork. Start a conversation, attach text files for extra context, or use voice input. Neuron remembers past conversations so you can pick one back up later.

Paste a requirement or user story and get back Functional, Negative, Boundary, and Accessibility test cases, grounded in real test-design principles rather than a generic checklist.

Pick a stack (Playwright, Cypress, WebdriverIO, or Selenium, across Java/Python/JS) and get a starter automation project: folder structure, config, setup commands, an example test, and best practices for that stack.

Describe a test scenario in plain language, pick from 8 language/framework combinations (Playwright TS/JS/Python, Selenium Java/C#/Python, REST API Python/Node.js), and get back a complete, runnable test.

Paste existing automation code and get an Issues / Suggestions / Best Practices review — anti-patterns, flakiness risk, and maintainability concerns, the way a senior SDET would flag them.

Give Olympus a user story, a live URL, or both, and its agents — one orchestrator plus requirements, planning, design, automation, real browser execution, defect triage, and reporting specialists — run the full cycle end to end. Execution is grounded in the real live page (not invented selectors) and runs non-destructively — it won't submit payments, delete data, or complete checkouts. If a step fails outright or a run finishes with failing test cases, click Retry (or ask in chat to "retest") to resume from there instead of starting over.

An autonomous Quality Engineer for mobile apps. Hand Pulse a native iOS or Android app and what you want proven — a written test case, or just a user story — and it does the rest. It learns the app, generates the test cases first when given only a user story, executes each step on a real device (resolving elements from the accessibility tree, on-screen text, and vision), validates the result against intent rather than brittle selectors, recovers from unexpected states like permission prompts and update dialogs, and returns an evidence-backed report with screenshots, video, and the reasoning behind every decision. The LLM decides what should happen; a deterministic engine (Appium / ADB / WebDriverAgent) performs it.