Replay documentation
CI & agents first
Record Playwright runs in CI, close the loop on failures from the PR, and give your coding agent full runtime context through Replay MCP. Everything here is built on the same deterministic, time-travel recordings that power Replay DevTools—we are doubling down on agentic workflows and less manual digging.
Run tests in Replay Browser and upload recordings from GitHub Actions—or wire up other CI providers.
Recent runs, top failing and flaky tests, and PR comments—visibility across your pipeline.
Connect recordings to Cursor, Claude Code, and other MCP clients so agents can inspect real runtime.
One-click recording from the browser when you are not driving everything from CI.
Not on GitHub Actions? Adapt the same install, env, and upload flow to your stack.
Playwright guide
Full setup: Replay browser, reporter, debugging, and FAQs.
PR comments
Root cause and context on the pull request when tests fail.
Top failing & flaky tests
Prioritize what breaks most often so agents and humans fix the right things.
MCP quickstart
Wire the server into your editor and start querying recordings.
Replay DevTools overview
Time-travel panels for console, network, sources—still the foundation under CI and MCP.
Curious how time travel fits in? Why time travel?