Ecosystem evidence
Real client code, checked end to end
Three exact SolidLil runtime surfaces, one archived partial LSX app, seven complete package apps, and four exact populated-package entrypoints run through reproducible behavior, size, speed, and memory gates.
Loading generated library results.
Interactive comparison
Start with shipped Brotli bytes
Choose a metric, filter the evidence boundary, and select exactly which rewrites to compare. Brotli-11 is the default and primary transfer-size view.
Loading comparisons.
Requested Vite 8 matrix
Exact published-entrypoint results only
Loading popular-library benchmark results.
Eligibility requires the complete selected public entrypoint, counted adapter code, no material representative time or retained-memory regression, and a LilScript Vite 8 artifact no larger than both npm/Vite 8 and public-API-preserving Closure ADVANCED in raw and configured-codec bytes. Size cells lead with Brotli-11, followed by gzip-9 and raw bytes. Partial implementations are not shown as library comparisons.
Passing exact entrypoints
Exact entrypoints under investigation
Candidate / partial ports (measured, not eligibility wins)
Same Brotli-11 / gzip-9 / raw cells. These rows are research ports or subset surfaces; they do not clear the exact-entrypoint publication gate. Every project name opens its Explorer detail page.
Openable Motion examples
motion@13 DOM fixtures, npm vs LilScript
Loading Motion lab examples.
Each fixture is a call-site Vite 8 bundle. Prefer the Demos dual-iframe stage for side-by-side playback. Size cells are Brotli-11 / gzip-9 / raw for the minified JS asset.
Audited, not claimed
Packages outside the compatibility boundary
A popular package stays out of measured totals when LilScript cannot express its complete documented root API. The blocker inventory is generated from the checked compatibility manifest.
Method
Measurements follow behavior gates
Loading benchmark metadata.
Transfer columns independently compress each served file with gzip level 9 and Brotli quality 11. Runtime is cache-busted Node module parsing plus deterministic app execution, not a browser rendering benchmark.