Low-code UI test automation platform powered by Playwright.
Create, record, and run browser tests through a web interface - no code required.
- Visual Recorder - click through your app via noVNC, steps captured automatically
- Smart Locators - uses
getByRole,getByLabel,hrefinstead of fragile CSS paths - Assertions Builder -
toBeVisible,toHaveText,toHaveURLand more - Device Presets - run desktop 1280px by default or emulate iPhone 15, Pixel 7, iPad and other devices
- Environments -
{{BASE_URL}},{{PASSWORD}}replaced at runtime per environment - Data-driven checks - define named scenario cases, run one selected case, or queue all enabled cases as a batch
- Scheduler - cron-based automatic runs with full history per schedule
- Suites - group tests and run them with one click or on schedule
- Trace Viewer - built-in Playwright trace viewer after every run
- Notifications - Telegram / Slack alerts on FAILED
- Export - download a single
.spec.tsor a runnable Playwright project.zip - Import - paste existing Playwright script, get a visual test
- Dashboard - pass rate over time and flaky test detection
| Feature | WrightTest | Cypress | Selenium IDE | Playwright UI |
|---|---|---|---|---|
| No-code recorder | β | β | β | β |
| Docker one-command | β | β | β | β |
| Mobile emulation | β | β | β | |
| Built-in scheduler | β | β | β | β |
Export to .spec.ts |
β | β | β | β |
Export runnable project .zip |
β | β | β | β |
| Self-hosted | β | β | β | β |
| Trace Viewer built-in | β | β | β | β |
Requirements: Docker, Docker Compose
git clone https://github.com/AlexFilippov-it/wrighttest.git
cd wrighttest
cp .env.example .env
# Edit .env - set JWT_SECRET to a long random string (required)
docker compose up --build| Service | URL |
|---|---|
| App | http://localhost:5173 |
| API | http://localhost:3000 |
| noVNC | http://localhost:6080 |
Default admin login is defined in .env:
ADMIN_EMAIL=admin@wrighttest.appADMIN_PASSWORD=changeme
On an empty database the seed also creates a Docker Demo project with two sample tests, a DEV environment, a Smoke Test suite, and an hourly schedule.
This path is the recommended first launch on any machine. The backend image is built on the Playwright-ready base image and includes the browser bundle, so no host browser or system library setup is required.
A public demo UI is available at:
Demo credentials:
- Email:
demo@wrighttest.com - Password:
demo
The demo account has read-only access only.
For a VPS or bare-metal deploy, make sure these are available first:
- Docker Engine
- Docker Compose V2 (
docker compose) - a reverse proxy such as nginx if you want a public domain
- free host ports for the published services if you are not proxying everything through nginx
Recommended Docker overrides for a server deploy:
FRONTEND_HOST_PORTBACKEND_HOST_PORTPOSTGRES_HOST_PORTREDIS_HOST_PORTVNC_HOST_PORTNOVNC_HOST_PORT
Keep container-internal ports unchanged; only adjust the host-facing ports and public URLs when needed.
For a server deployment:
- Copy
.env.exampleto.env - Set
JWT_SECRETto a long random value - Set
FRONTEND_URLto your public domain or reverse-proxy URL - Set
VITE_BACKEND_URLandVITE_NOVNC_URLto the URLs users should reach from the browser- local example:
http://localhost:6080 - reverse-proxy example:
/vnc
- local example:
- Adjust
*_HOST_PORTvalues if your VPS already uses ports like3000,5432, or6379 - Run
docker compose up --build -d
If you place nginx in front of the app, proxy the public frontend domain to the frontend container and route backend/API traffic to the backend container. The backend already exposes /health and /health/db for readiness checks.
WrightTest uses a single VITE_NOVNC_URL value and derives the websocket path from it at runtime.
Use this when noVNC is exposed directly on localhost:
VITE_NOVNC_URL=http://localhost:6080The recorder iframe connects to:
ws://localhost:6080/websockify
Use this when nginx exposes noVNC under a path prefix:
VITE_NOVNC_URL=/vncThe recorder iframe connects to:
wss://your-domain.example/vnc/websockify
This keeps the container image unchanged and moves the deployment-specific part into configuration only.
If you are working with an AI coding agent, start here first:
It contains the canonical repo workflow, startup order, and environment rules for WrightTest.
Use this only if you want to run the frontend with Vite and the backend on the host.
Requirements: Node.js 20+, PostgreSQL 16, Redis 7
git clone https://github.com/AlexFilippov-it/wrighttest.git
cd wrighttest
cp .env.example .env
# Edit DATABASE_URL for local PostgreSQL
npm install
npm run setup
cd backend && npx prisma migrate dev && npx prisma db seed && cd ..
npm run devnpm install now runs a Playwright bootstrap step for Chromium and WebKit. On Ubuntu/Linux
it will also try to install system dependencies when the terminal session allows
it. If Playwright still cannot launch Chromium or mobile recording needs WebKit,
run the same bootstrap manually and then install Linux deps:
npm run setup
npx playwright install chromium webkit
sudo npx playwright install-deps chromium webkit| Service | URL |
|---|---|
| Frontend | http://localhost:5173 |
| Backend | http://localhost:3000 |
If the host environment still reports missing Playwright libraries, rerun npm run setup once from the repo root. On Ubuntu/Linux this may fall back to npx playwright install-deps chromium webkit when needed.
git pull
docker compose up --build -dOn first launch or after resetting volumes:
cp .env.example .env
# Make sure JWT_SECRET is set to a long random value
docker compose up --build -dMigrations apply automatically on startup. Existing projects, tests and run history are preserved in the Postgres volume.
All ports are in .env - no hardcoded values in code:
BACKEND_PORT=3001
FRONTEND_PORT=8080
NOVNC_PORT=6081Then restart:
docker compose up --build -dProjects overview with health summaries, project status, and onboarding.
Project workspace with checks, schedules, alerts, and operational summaries.
Edit browser checks visually and review global execution history across projects.
Live recording captures Playwright-ready selectors directly from the browser session.
WrightTest separates checks from test cases:
- Check - one browser scenario: URL, device, and steps.
- Test case - one named set of scenario variables for that check.
- Run - one execution of one check with one selected test case.
- Run batch - a grouped set of runs created by running all enabled test cases for one check.
This means a project can show Checks: 1 while that check contains multiple test cases. The check is counted once because the browser flow and steps are shared; each enabled case creates its own run when executed in a batch.
Each test case has:
- a human-readable name
- an enabled/disabled state
- scenario variables such as
EMAIL,PASS,EXPECTED_MESSAGE, orEXPECTED_URL
All variables are equal: they can describe form input, expected results, URLs, titles, search text, order numbers, or any other scenario value. Use them in the Start URL or step fields with the same {{VARIABLE}} syntax used by environments.
Example case:
Wrong password
EMAIL=admin@test.com
PASS=123456
EXPECTED_MESSAGE=Invalid email or password
EXPECTED_URL=https://demo.wrighttest.com/login
Example steps:
Fill {{EMAIL}}
Fill {{PASS}}
Assert text {{EXPECTED_MESSAGE}}
Assert URL {{EXPECTED_URL}}
For manual runs in the editor, select the case in Check settings next to the Environment selector, then run the check. If more than one case is enabled, use Run all enabled cases to queue a batch. Each case in that batch creates a separate TestRun, keeps its own variable snapshot, screenshots, trace, error, and step results, and runs through the existing worker.
On the project checks page, Run is data-aware: checks with one enabled case start a normal run; checks with multiple enabled cases queue a batch and open the batch result page.
WrightTest combines variables from the selected environment with variables from the selected test data case for that run. Test data case variables take precedence when the same variable name exists in both places. Empty strings are valid values, so a case can intentionally define EMAIL=.
Disabled cases are ignored by run actions and do not block variable diagnostics.
This keeps ordinary checks unchanged: if a check has no test data, it runs exactly as before.
If no device is selected, WrightTest uses the default desktop browser context (1280x720). The device selector only stores explicit overrides such as:
Desktop 1280pxDesktop 1920px (HiDPI)- mobile and tablet presets from Playwright
There is no separate saved value for "Desktop default"; leaving the selector empty is the default desktop mode.
Create named data cases in the check editor, then select the enabled case to use for a manual run.
Export a complete runnable Playwright project as a .zip archive.
The generated project includes:
- Playwright configuration
- ready-to-run test files
package.json- optional environment variable support
- minimal project structure for local IDE usage
After extraction:
npm install
npx playwright install
npx playwright testExport a runnable Playwright workspace that opens directly in your IDE and runs locally without manual setup.
WrightTest supports project-level access roles to control who can edit, run, and manage a project.
Editors can:
- create and edit checks
- run checks
- use recording
- manage schedules and environments
- export Playwright specs and projects
- view runs, screenshots, and traces
Viewers have read-only access.
Viewers can:
- view checks and steps
- inspect runs and results
- open screenshots and traces
- explore project structure
Viewers cannot:
- modify checks
- run checks
- start recording
- edit environments or schedules
- manage project members
Project-level roles keep Viewer accounts read-only while Editors can manage checks, schedules, environments, and exports.
βββββββββββββββ POST /recordings/start ββββββββββββββββββββ
β noVNC β ββββββββββββββββββββββββββΆ β playwright codegenβ
β (iframe) β βββββ sessionId βββββββββ β headed browser β
βββββββββββββββ ββββββββββββββββββββ
β clicks recorded as Steps
βΌ
βββββββββββββββ POST /tests/:id/run ββββββββββββββββββββ
β Step Editor β ββββββββββββββββββββββββββΆ β BullMQ + Redis β
β + Validate β β Worker queue β
βββββββββββββββ ββββββββββββββββββββ
β
Playwright headless
β
βββββββββββββββ polling GET /runs/:id ββββββββββββββββββββ
β Run Result β ββββββββββββββββββββββββββ β Screenshots β
β Trace Viewerβ β Traces β
βββββββββββββββ ββββββββββββββββββββ
| Layer | Technology |
|---|---|
| Frontend | React + TypeScript + Vite + Ant Design |
| Backend | Node.js + Fastify + TypeScript |
| ORM | Prisma + PostgreSQL |
| Queue | BullMQ + Redis |
| Runner | Playwright (Chromium) |
| VNC | noVNC + Xvfb + x11vnc |
| Auth | JWT + bcrypt |
| Container | Docker Compose |
- Network mocking (
page.route()) - CLI tool (
wrighttest run --project-id) - Export full Playwright project
- Test-to-Doc export
- Allure / TestIT integration
The Docker image badge will be added after the first public image publish.
WrightTest is source-available, but not open-source under the OSI definition.
You may use, copy, modify, and run WrightTest for personal, educational, research, internal, and evaluation purposes, including testing your own applications, websites, services, or products.
You may not sell WrightTest as a standalone product or offer WrightTest, or a modified version of WrightTest, as a public hosted service without prior written permission.
See LICENSE for details.









