wtf is playrun?

PlayRun is a tool for running large collections of Playwright browser tests on Kubernetes, with a visual interface for planning and watching those runs. Instead of a single long test script, you build a flow graph in the app: a smoke test that gates whether the rest should run, fanning out into per area test suites, all converging into a final notification step. A server side controller then turns that graph into Kubernetes jobs and streams the results back into the same app you built the flow in, so there is no separate reporting server to maintain. Under the hood it is a Next.js web application backed by a PostgreSQL database, with test results and recordings stored in any S3 compatible storage service. It talks to a Kubernetes cluster directly to schedule one pod per test run, and it supports logging in through Okta today, with more login providers planned. You can try the whole interface without installing anything beyond Docker: one command starts the app in a simulation mode where the UI and run history work against fake test progress, useful for exploring what it looks like. To see it schedule real Kubernetes pods and run actual Playwright tests on your own machine, a separate script sets up a small local cluster along with its own database and storage, using a public example test suite to demonstrate the flow. For a real deployment, PlayRun reads which tests exist directly from your Playwright repository on GitHub, authenticating with a personal token, an app installation, or nothing at all for public repositories. Configuration for things like the container image, test tags, and infrastructure settings lives in a single file that reloads automatically when changed, while secrets are kept separately in environment variables or your own secret manager. The project uses a source available license: you can use, modify, and self-host it freely, but you cannot resell it as a competing hosted product until two years after each release, at which point that version becomes fully open source.