Each one started as a script inside my own work, then got cleaned up, tested and published.
Python CLI · MIT
wordcaps
Word-by-word karaoke captions for any video, made on your machine.
$ wordcaps burn talk.mp4 --lang en
Transcribes with Whisper, then treats timing as the real problem: silence-aware transcription, word edges snapped to the loudness envelope, and a coverage check that re-reads the finished file and fails if any speech has no caption.
How findable is a website on Google and in AI answer engines?
$ npx geocheck example.com
Audits crawl and index, on-page SEO, AI readiness (10 AI crawlers, JSON-LD entities, quotable answer blocks, llms.txt) and local signals. Every issue is a measurement with the page it was measured on. Zero runtime dependencies.
Opens on today's workout with last session's weights already filled in. Supersets, history, progress charts and a nudge when it is time to add weight. Works offline; the data never leaves the phone.