Using VoiceFlow

History & stats

Every transcription is saved locally as JSON. Browse, search, re-paste, export, and see aggregate stats — all on your machine.

Format

One JSON file per record

Location

userData/history/

Export

JSON or Markdown

Browsing history

The History screen lists records newest-first, paginated. You can search across the final text, original transcript, app name, and window title, and filter by mode (all, dictation, or ask).

  • Searchmatches a substring across the record's text and captured app context.
  • Filter narrows to dictation-only or ask-only records.
  • Records are cached in memory and re-sorted on each save so the list stays fast.

Re-paste, export, delete

Re-inject

Paste a saved result back into the current app, using the same paste behavior as the original mode.

Export

Save a single record or the whole history to disk. Use a .json or .md extension to pick the format.

Delete

Remove a record; its JSON file is deleted from disk and the cache is updated.

Aggregate stats

VoiceFlow computes running totals across every history file:

StatTypeHow it's computed
totalWordsnumberSum of per-record word counts. Counting is CJK-aware — each CJK character counts as one word, Latin words split on whitespace.
totalCountnumberNumber of saved transcriptions.
totalDurationSecondsnumberSum of recording durations, measured from recording start to stop.

Where it lives

History files are stored under Electron's userData directory in a history/ folder. You can point VoiceFlow at a different directory (for example a synced folder) and it will move future reads and writes there. For the exact record shape, see Data model.