Plan sprints, run live QA sessions in an embedded browser, drive a 12-tool automation hub with self-healing tests, generate cases with AI, and sync everything to OpenProject — one workspace, end to end.
$ qa-os session start --project checkout --live
▸ Launching embedded browser (CDP stream attached)
▸ Recording session video… ● REC
✓ Step 1 · Open /checkout — 1.2s
✓ Step 2 · Add item to cart — 0.8s
✗ Step 3 · Apply coupon "SAVE10" — assertion failed
▸ Bug drafted → BUG-112 "Coupon not applied to total"
▸ Syncing BUG-112 to your tracker… work package #4821 ✓
$
28+
platform modules
12
automation hub tools
4
AI providers — incl. free local
3
trackers synced (OpenProject · Jira · Azure)
// 01 · plan & manage
Everything upstream of the first test: scrum ceremonies, a real requirements hierarchy, business requirement documents, and analytics that read from live data.
Jira-style sprint planning: drag-and-drop backlog, kanban board, and a full task drawer with subtasks, work logs, attachments, watchers, and votes.
A real Epic → Feature → Story hierarchy with workflow transitions, sprint start/complete, and a requirements kanban board alongside it.
Upload, version, and preview business requirement documents (PDF + Word), edit them in a Word-style editor with docx export, and discuss sections with @mention comments.
A live analytics dashboard: KPI tiles with sparklines, burndown, velocity, cumulative flow, defect trends, team radar, and a coverage funnel.
Two-tier access control — global admins plus per-project Owner, Admin, Member, and Viewer roles — with token-based email invitations.
The command dashboard: cross-project health, test and bug KPIs, and recent activity in one screen.
// 02 · manual testing
A live embedded browser, structured test cases, and a bug lifecycle built for retesting — every finding traceable back to the session, the step, and the story.
Test inside an embedded browser streamed straight into the app. The whole session is video-recorded, and any moment becomes a bug with one click — annotated with arrows, highlights, and blur before it's filed.
Every ended session preserved with its recordings, screenshots, and the bugs it produced — each bug links back to the exact session it came from.
Manual and automated suites with priorities, structured steps, tags, bulk operations, and full execution history.
Defects move through Open, In Progress, Ready for Retest, Reopened, and Closed — with severity, source, screenshots, repro steps, and bulk triage actions.
Click through your app in a real browser and replay the recording as an automated test on the local Playwright worker.
Requirements Traceability as a live grid — every story linked to the tests that verify it, with untested requirements surfaced instantly.
// 03 · automation hub
The Automation Hub wraps the execution engine in a full toolchain — from authoring and evidence to AI diagnosis, self-healing, flake analysis, and cross-browser scale.
Author and launch automated suites with test data, accessibility checks, and execution targets.
Visual flow recorder that turns real browser sessions into repeatable automated tests.
Evidence Vault: screenshots, animated GIFs with a frame viewer and highlighter, plus live-streamed execution logs.
Execution traces unpacked into timeline, network, console, and observability views.
AI failure diagnosis with a full diagnostics bundle — trace, environment metadata, and a redaction scan.
Self-healing locators with a locator-health report and estimated resolution time per broken selector.
Flaky-test detection plus a wait-analysis profiler that flags reducible waits from real trace timings.
Local cross-browser grid plus Cloud & Parallel: BrowserStack config, Docker runs, and sequential-vs-parallel savings.
Pixel-level baseline comparisons with a baseline manager, comparison runs, and tunable thresholds.
Execution videos with trace-action markers, jump-to-failure, and ffmpeg clip export.
Chat with your suite: command library, per-provider setup, real token and cost tracking, consent and budgets.
Engine properties editor, environment variables, AES-GCM encrypted secrets with an audit log, and live system logs.
// 04 · ai copilots
Generation, diagnosis, and design comparison — powered by Ollama running locally, Hugging Face, or OpenAI, with a free rule-based mode that needs no key at all.
Generate structured test cases from your requirements — or straight from an uploaded document.
An agent workspace that reasons over your project — requirements, tests, and bugs — to answer questions and draft work.
Compare the live UI against your Figma designs and catch drift before users do.
Point at any element in a live page and get a stable selector for your tests.
Central datasets for your suites — shared between manual runs and the automation engine.
Author flows once, reuse them across test cases as shared building blocks.
// 05 · collaboration & sync
Chat, notifications, and profiles built in — and a sync layer that meets your tracker where it already lives.
Real-time channels and private DMs, scoped to your workspace — triage without leaving the app.
A real-time bell wired into everything: assignments, mentions, sync results, and run outcomes.
Public member profiles with activity history, plus per-project teams and roles.
Two-way sync pushes bugs and stories to OpenProject as work packages — with connection health, key mapping, and idempotent board placement. Team members sync from Jira, Azure DevOps, and OpenProject too.
// the pipeline
Create an account and run your first live QA session in minutes.