How-to guides

Short, task-focused guides for using Voxint day to day. New here? Start with Setup to install it, then the first-run walkthrough (setup wizard + guided tutorial).

Guide	For when you want to…
Add media & manage runs	Get a recording into Voxint (upload, URL, or a watched folder), follow it through the pipeline, requeue, cancel, or archive a run, and search across your transcripts by meaning.
Review & adjudicate	Confirm who each speaker is and correct the transcript: the review queue, the workbench, keyboard shortcuts, the waveform, and splitting or reassigning segments.
Manage speakers & export	Tend the speaker roster (rename, merge, archive), read a finished transcript on screen, and download it (txt / md / srt / vtt / json / rttm).
Translate transcripts	Turn a finished transcript into another language with your configured LLM, read the translation beneath each original line, and download it in every transcript format.
Settings & troubleshooting	Configure everything from the browser, check what your hardware is doing, and fix common problems.
Checking for AI-generated speech	Turn on synthetic-speech detection, score a recording (automatically or by hand), read the risk levels, and understand the known limitations.
Changing pipeline models	Advanced: point transcription or diarization at a model other than the validated default, what it costs, and how to apply the change.

Reference material (architecture, the model-service contracts, operations, and the release process) lives one level up in docs/.