Subtitle reading-speed lab: a dependency-free SRT text-count and cue-duration example

An AI-assisted, MIT-licensed educational source example for comparing text counts and display durations in basic SRT cues. The archive includes the complete browser example, English and Chinese usage notes, an MIT licence and a dependency-free Node.js logic-check suite. Synthetic cues illustrate crowded text, CJK characters, joined emoji and invalid durations. It parses basic SRT timestamps, normalises whitespace and common formatting, and uses Unicode grapheme clusters where Intl.Segmenter is available with a documented code-point fallback. The chosen threshold is illustrative and user-controlled, not a universal reading or accessibility standard. No transcription, translation accuracy, human-participant evaluation, benchmark or production-data claim is made. The calculation makes no network calls. An earlier web demonstration was accepted on 21 October 2026; this archival edition preserves the existing source and includes additional documentation, rather than claiming a first-ever release. Maintained by Scanfile contributors with AI assistance. Affiliation: Scanfile. The separate translation service is not the source code represented by this archive. Affiliation website: Scanfile