What's new in v2.0

Version 2.0 is a major update to benchmarking and the REPL. It expands the kinds of code you can test, makes results more useful, and gives you better tools to inspect and share each run.

Benchmarking

Benchmarking now uses a new measurement engine, provides richer statistics, and includes better controls for longer or more detailed runs. The highlight is the entirely new, long-requested DOM mode: you can finally measure code that uses real browser APIs and HTML, while Worker mode remains available for pure JavaScript/TypeScript tests.

  • Run DOM benchmarks with browser APIs and an optional HTML fixture. Every case gets a fresh sandboxed frame.
  • Choose Quick, Standard, or Extended run lengths, then run cases sequentially or in parallel.
  • Measure calibrated batches after a dedicated warmup and track live progress during the run.
  • Compare median, p95, standard deviation, 95% RME, and sample counts alongside ops/s.
  • Export a PNG or copy results as Markdown, CSV, or JSON. PNG exports can include statistics and optional device and browser details.
  • Keep run controls available while scrolling, and follow the system theme or choose a light or dark theme.

REPL

The REPL has grown from a simple result viewer into a more complete environment for browser experiments. It now supports DOM previews and shows the details of a run as they happen instead of only returning the final value.

  • Choose Worker or DOM mode. DOM mode adds an HTML fixture and a fresh sandboxed preview.
  • Stream results, console output, errors, and timing markers.
  • Copy the latest results/investigation as Markdown.