Playwright Reports Server

A self-hosted home for your Playwright test results. Stores reports, merges sharded runs, tracks flakiness over time, and (optionally) runs LLM failure analysis you can read from the UI or pipe into your code assistant.

What you get:

• Reports & results storage: keep HTML reports in one place, view them without downloading.
• Sharded run merging: server-side equivalent of playwright merge-reports, without scripting the post-step.
• GitHub Sync: pull Playwright artifacts from GitHub Actions on a cron without touching workflow files.
• Overview and Analytics dashboards: statuses, pass-rate trends, slow tests, failure categories, failure clusters.
• Test flakiness & quarantine: score test stability, optionally skip the noisiest.
• LLM failure analysis: pick any OpenAI-compatible or Anthropic-format provider; analyses surface inline on each report and on the dashboard. Bring your own token casino.
• Code-agent integration: the pwrs-cli plus Claude Code skill exposes failure context to Claude Code, Codex, Cursor, and the LLM-driven code assistant that drops next week.
• Notifications: send report details and summaries to Slack or any webhook. Event rules on upload, schedule rules on cron, delivery log. See notifications docs.
• Pluggable storage: local filesystem, S3-compatible, or Azure Blob, see storage docs.
• Accounts, roles, and SSO: optional auth, off by default; set API_TOKEN to turn it on. Cookie sessions with scrypt-hashed passwords, three roles (admin / reader / readonly), invite-only onboarding, scoped API keys for CI and the CLI, and optional single sign-on via GitHub, Google, or any OIDC provider (Okta and friends). See Authentication.
• Persistent state with SQLite and Litestream replication for S3 and Azure Blob storages. See storage.

Demo

demo-playwright-reports-server.koyeb.app

Getting started

Prerequisites

• Node.js v22+
• pnpm v9+

Install & run

git clone https://github.com/CyborgTests/playwright-reports-server.git
cd playwright-reports-server
pnpm install

Production:

pnpm run build && pnpm run start

Server listens on port 3001 by default (configurable via PORT) and serves both the API and the built SPA from one process. Open http://localhost:3001.

Development:

pnpm run dev

Backend on :3001, Vite dev frontend on :3000. The frontend proxies /api/* to :3001, so during dev you can hit either port for the UI.

All persistent state lives in apps/backend/data/ for pnpm run start and dev, or /app/data/ in the Docker image. Back this up to keep reports, results, and the SQLite metadata file safe. Or don't.

Docker

docker run -p 3001:3001 -v /path/on/host:/app/data \
  ghcr.io/cyborgtests/playwright-reports-server:latest

Full deployment options: see Deployment.

Documentation

The detailed docs live in the docs site:

Topic	Page
Every env var	Configuration
Accounts, roles, sessions, API keys, SSO	Authentication
fs / s3 / azure backends and Litestream replication	Storage
Title, logo, favicon, header links	White‐label
Deployment	Deployment
How to upload reports	Uploading Reports
Analytics stats, trends, failure clustering strategies	Analytics
Flakiness scoring, quarantine, reporter-side skip	Test Management
LLM analysis	LLM Analysis
Claude Code / Codex / Cursor integration via pwrs-cli	Code Assistant Integration
Slack & webhook notifications, event + schedule rules, templates, delivery log	Notifications