This is the provenance record. It shows when this first appeared in my log, what happened to it afterward, and how each of those claims can be checked.
The in-browser Web Speech API barely works on iPhone/Safari, so the assistant "couldn't hear" reliably. Switch to recording the command with MediaRecorder and transcribing it server-side with Whisper — works the same on iOS and desktop.
Independently witnessed. This build is tied to GitHub pull request #6188, merged at the timestamp above. That timestamp was written by GitHub, not by this site, and I cannot alter it.
Timestamped by GitHub, not by this site. The repository is private, so this is shown on request rather than as a public link.
This record was added since the last seal and will be included in the next one.
Permalink: https://jeffreylong.com/proof/e/501