What's built, and
what isn't.

Answer quality, measured rather than guessed

Troctor used to refuse to answer below a hand-picked confidence floor. That floor was removed in v0.1 because on a small vault it refused constantly, which is worse for a new user than a hedged answer that cites what little it found. The honest position is that we never knew whether the number was in the right place, because measuring it means testing against real private transcripts, which by design we cannot see. Building an evaluation harness that runs on your machine and reports only a score is the first thing on this list, and a measured floor is what it would buy back.

Live mode: in daily beta use

This is the feature people ask about first, and all of it now exists in the repository. It captures what the other people say, streams it to Soniox for transcription while the meeting runs, tells speakers apart with diarisation so each voice is its own named block, writes the transcript into that meeting, and answers questions the room asks out loud, under the line that asked them. An eye icon on any speaker chooses whose questions are answered automatically, everyone else remains a highlight away, and each answer knows what the ones before it already said, so an interview does not hear the same story twice. The capture is system audio through a Core Audio process tap; your own microphone is not captured at all.

This entry used to say it has never been run, then that real sessions had happened. It is further along now: beta users run it through real meetings daily, and the faults those meetings exposed have each been fixed and pinned by a check, from answer pacing tuned for interviews to a transcript label that ran into the text it labelled. A device change no longer ends a meeting, and a recording is written continuously to your own disk so a crash cannot take a call with it.

It stays on this page because it is still moving quickly: the live path changes most weeks in response to what beta meetings surface, and a feature being reshaped that often is not one we would ask you to build a habit on yet. The honest description is in daily use and still being hardened, rather than finished.

Four things about it that will not change, and which are on the security page in full. It captures nothing unless you switch it on for a particular meeting. The audio goes from your Mac straight to Soniox and never through us, on our account and our bill, with no daily limit during the beta. It writes a recording to your own disk, at about 115 MB per hour, which you can reveal, delete or switch off. And it sends the audio of the meeting to a company that is not us.

Watching a folder

Right now you import by hand. If your existing tool already drops transcripts into a folder, Troctor should notice and index them without being asked. Small, useful, and the thing that turns this from a tool you remember to use into one that is simply current.

More sources, properly parsed

Every export format has its own idea of what a speaker label looks like. The parsers handle the common cases and fail politely on the rest. Each new source is a small amount of work and a large amount of testing, and we would rather add them in response to a file that broke than in anticipation.

Windows

Asked for, and not soon. The interface is Electron and would port. The storage layer, the credential handling and the packaging would all need doing again. macOS has to be genuinely good before that is a sensible use of the only pair of hands here.