Skip to content

Interactive Playground

Experiment with Branchline directly in your browser—no installs required. Pick an example, edit the program and input, enable tracing, and run. The catalog can be filtered by example category and AI subset compatibility so docs, tests, and AI-authoring examples stay discoverable from the same descriptor metadata.

Tips: Cmd/Ctrl + Enter runs. Switch JSON/XML input. Toggle tracing for EXPLAIN(...). Use Inspect for normalized source, diagnostics, and AI subset blockers. Changing examples resets the editors.
Open in new tab Use ?example=<id> in the URL to preload an example (e.g., ?example=stdlib-hof-overview).

Curated examples (all preserved)

Use the category and AI subset filters in the embedded playground to narrow this list. Examples marked AI compatible are intended for inspect-first authoring loops; examples marked incompatible are still runnable language examples, but Inspect will explain why they are outside the AI canonical subset.

Troubleshooting

  • If the embed does not appear, open the new-tab link above.
  • XML inputs are parsed into objects; attributes use the @attr convention and text nodes use #text.
  • The playground runs entirely in your browser using the Kotlin/JS interpreter; no data leaves your machine.

Contract JSON in Playground

  • Enable Contracts to inspect inferred input/output contracts.
  • Keep debug off for clean JSON intended for docs/tooling.
  • Enable Contract debug when you need origin and available span metadata.

Inspect workbench

  • Run an example and use Inspect → Normalized source to copy the canonical form for docs, agents, and repair loops.
  • Use Diagnostics for parser, contract, and runtime-facing messages before executing a changed program repeatedly.
  • Use Subset blockers when an example is runnable Branchline but not AI-canonical; blockers are driven by the same inspect compatibility result used by CLI automation.