QA QASyn / User guide
Autonomous quality engineering

Everything you need to plan, run and prove quality.

QASyn turns a crawl, a requirement or a pull request into executable browser tests — then keeps a screenshot-backed record of every step. This guide adapts to your role, lets you search all 19 modules, and walks through the six workflows teams use most.

Start a walkthrough →
19product modules
11test types
6guided walkthroughs
100%step screenshot evidence
Start here

Choose your role

Permissions in QASyn are role-based. Pick a role to see the surfaces and actions available to it — the feature explorer below follows your choice.

All roles

A shared tour of the workspace. Use the tabs to focus the guide on what your role can do.

  • Navigate the 19 modules from the left sidebar
  • Open this guide any time from Help & docs
  • Use the search below to find a feature by name or outcome
Permission note Admins manage billing, team and settings. QA authors and runs tests. Engineers use Studio, API and integrations. Eng managers read reporting and PR risk. Super Admin provisions tenants across workspaces.
Feature explorer

Find a capability

Search by module, action or outcome. Cards are tagged with the roles that use them most.

25 features
Try “screenshot”, “Bivia”, “enterprise” or “regression”.

Overview

Quality KPIs, Bivia token observatory and the Quality Autopilot launcher in one landing page.

Module 01All roles

Pull requests

Impact analysis, risk scoring, provider comment, CI/CD gate Markdown and auto-regression on the deployed environment.

Module 02RegressionEng manager

Crawl

Map a site, record flows into requirements and generate linked draft cases. Re-import updates without duplicates.

Module 03Autopilot

Code Analysis

Graphify topology, blast radius and coverage gaps so you know which code a change can break.

Module 04Bivia

Requirements

Import specs, track coverage, link test cases and export the library with per-column filters.

Module 05Traceability

Test Cases

The master grid: all 11 test types, folders, step groups, bulk run/delete/export and per-column filters.

Module 0611 typesQA

Step Groups

Reusable multi-step building blocks shared by test cases; edit once and every consumer stays in sync.

Module 07Reuse

Test Studio

Author NLP steps, AI assertions and variables. Ground a selector, self-heal it, inspect versions and see the screenshot for every run.

Module 08ScreenshotsQA

Test Suites

Group cases by platform and module, track the latest result and run or export the suite in bulk.

Module 09Admin

Test Plans

Combine suites, bind environments, schedule runs, compare run matrices and export reports.

Module 10SchedulingEng manager

Environment & Data

Environments, test-data profiles, global variables and auth profiles with encrypted secrets and CSV import.

Module 11Admin

Uploads

Upload specs and spreadsheets, parse them into requirements and convert them into draft test cases.

Module 12Import

Knowledge Base

Indexed product knowledge that improves AI generation and gives Bivia the context it needs.

Module 13Context

Team & access

Invite members, assign roles, verify accounts and control what each person can change.

Module 14Admin

Tunnels

Reach localhost or protected networks through a secure connector when no staging URL exists.

Module 15Engineer

Integrations

Connect source control, trackers, chat, Bivia and browser providers with encrypted credentials.

Module 16Admin

Billing & usage

Plan, credits, ledger and the 1B-token Bivia observatory with calibration and abstention metrics.

Module 17GovernanceAdmin

MCP

Query your workspace from an AI assistant with a scoped API token and the Model Context Protocol.

Module 18Engineer

Settings

Organisation profile, audits, analysis defaults, bug reproduction, API tokens and the Super-Admin Console.

Module 19Super Admin

Per-column filters

Every data table header has a filter. Categorical columns offer distinct values; dates and text use substring search. Active filters show a dot and clear in one click.

TablesAll roles

Screenshot evidence

Every executed step stores a screenshot in the step and run log. Step Details, run history and self-heal all render the persisted image.

EvidenceAll roles

Bivia grounding

Resolve an element from a description, DOM and screenshot. Get a calibrated selector, confidence and a 0–1000 bounding box.

BiviaQA

Self-healing

Detect corrupt selectors, ask Bivia for a replacement, persist it with a version entry and re-run the case with fresh evidence.

BiviaEngineer

Super Admin provisioning

Create an enterprise tenant with environments, a credit grant and a verified admin, then configure plan, seats, SSO and SLA.

Super AdminSuper Admin

REST API

Every dashboard action is backed by an authenticated route handler. Use the API drawer for the endpoints behind the six walkthroughs.

APIEngineer
Hands-on

Six walkthroughs

Each walkthrough maps a real user goal to the screens and API calls that complete it. Expand one to follow the steps.

1 Author, type and run your first test case ⌄

Goal: go from a blank workspace to a passing run with screenshot evidence on every step.

  1. Open Test Cases and choose New Test Case. Pick AI Assistance, Import, Record or Manual.
  2. Select one of the 11 test types in the Test Type field (Functional is the default), choose the module and priority, then create.
  3. Open the case in Test Studio. Confirm the Navigate step points at your real URL and add Click, Enter Text or AI Assertion steps.
  4. Click Run. QASyn resolves the target URL, executes each step, and writes a scoped TestCaseRun.
  5. Open Step Details or Run history and confirm the screenshot thumbnail is present for every executed step.
  6. Click any thumbnail to open the Visual Evidence Lightbox: status, MIME badge, timestamp, 🔀 Compare with Previous Run time travel and ⬇ Download Evidence.
Evidence is persisted twice: on stepsJson[i].screenshotBase64 and in the run’s logJson. The Type column always shows the exact type you chose.
2 Ground a selector and recover it with self-heal ⌄

Goal: fix a broken or empty selector without guessing CSS.

  1. In Test Studio, select the failing step and open Step Details.
  2. Click Run Bivia Visual Grounding. QASyn uses the step URL, then variables.url / targetUrl / baseUrl, then the first Navigate target, then the org website.
  3. Review the returned bounding box, confidence and screenshot.
  4. Click Simulate Self-Healing. Bivia proposes a new selector; QASyn persists it on the step, writes a version entry and creates an “Auto-fix rerun”.
  5. Confirm the healed screenshot card and the updated run history show the new evidence.
  6. Click the healed or step thumbnail to open the Visual Evidence Lightbox and compare the previous vs latest run side by side with the opacity slider.
If the target site is unreachable, grounding still returns generated visual evidence and explains whether the DOM/screenshot was live or synthetic.
3 Turn a crawl into a regression suite ⌄

Goal: build a reusable baseline from a deployed application.

  1. Open Crawl and launch a crawl for the target URL (or use Quality Autopilot from Overview/Crawl for the full five-phase flow).
  2. Wait for the crawl to complete. QASyn records flows as requirements and generates linked draft test cases.
  3. Open Test Cases, filter the Type or Module columns, and review the generated cases.
  4. Create a Test Suite, add the cases, then create a Test Plan that binds the suite to an environment and schedule.
  5. Run the plan and inspect the run matrix; every case keeps its own per-step evidence.
Re-importing the same crawl updates the existing linked cases instead of creating duplicates.
4 Gate a pull request and run auto-regression ⌄

Goal: decide whether a PR is safe to merge using impacted tests and a deployed environment.

  1. Open Pull requests and analyse a PR (or open an existing row). Changed files are mapped to modules and impacted cases.
  2. In the drawer, click Generate CI/CD PR Gate Comment. Review the verdict, Bivia risk score, impacted test table and Strix security summary, then copy the Markdown.
  3. In the PR auto-regression section, choose the environment (Production, Staging or any configured environment) or override the deployed URL.
  4. Keep Auto-create PR-specific test cases checked. QASyn creates a Smoke and a Regression case named PR #n · … if none exist.
  5. Click Run PR auto-regression. The result panel shows pass/fail/regressions, the target URL and the created cases; the PR row updates its verdict and counts.
A caught regression sets the PR to regression_caught. Open the affected case → run history to inspect the failing step’s screenshot before merging.
5 Provision and configure an enterprise tenant ⌄

Goal: onboard a new customer without touching SQL or the database console.

  1. Sign in as the super admin and open Settings → Organization → Super-Admin Console (or use the workspace menu).
  2. Click Provision enterprise. Fill in organisation, website, admin name/e-mail/password, seats, environment name/URL and credits.
  3. Submit. In one transaction QASyn creates the Enterprise organisation, Production + Staging environments, a credit ledger grant, a verified Admin user and an audit entry.
  4. Use the per-organisation Configure action to change plan, grant credits, update seats/environment and toggle SSO/SLA.
  5. Switch into the new workspace from the sidebar to verify its environments, billing and team.
Enterprise defaults to 250,000 credits. The ledger grant is visible in Billing & usage and the admin can sign in immediately.
6 Filter, export and operate with confidence ⌄

Goal: find the exact rows you need and share evidence with the team.

  1. On any data grid (Test Cases, Crawl, Requirements, Suites, Plans, Team, Audits, Pull Requests), click the filter icon in a column header.
  2. Categorical columns such as Type, Status or Priority offer a distinct-value dropdown; dates and text accept a substring search.
  3. The filter icon shows a blue dot when active. Reopen it and choose Clear, or use the toolbar filter chips, to reset.
  4. Select rows to reveal the bulk bar: run selected, export CSV or delete (permission-dependent).
  5. For operational context, open Billing & usage for credits and the Bivia observatory, Integrations for providers, or MCP to query the workspace from an assistant.
Existing search, column visibility, pagination and RowMenu actions stay available — per-column filters are additive.
At a glance

What every run guarantees

Real URL grounding

The test case’s real target wins over the org default every time.

Step-level evidence

Passed, failed and healed steps all persist screenshots.

Deterministic PR runs

Auto-regression uses stable outcomes so PR verdicts do not flip randomly.