Skip to content
How to · Mac voice writing

Is SuprFlow's Voice to text mode really a raw transcript?

SuprFlow 1.0.0 does not show a Voice to text preset in Formatting, and its Default mode still applies local cleanup. See what that means if you need a verbatim transcript.

SuprFlow 1.0.0 contains a Voice to text preset, but we found no user-facing control that activates it. The installed app 1.0.0 (12) shows Default, Cloud API, and Local (Ollama) on its Formatting screen, without a Voice to text or Verbatim control there. If the preset is activated in another build, it forces the optional formatting provider to None. That avoids a selected Cloud API or Ollama formatter for that path. It does not, by itself, mean the inserted words are byte-for-byte raw speech recognition output. When the formatting pipeline is enabled, the app still runs a local cleanup pass with no provider selected, which can alter punctuation, capitalization, or filler words before insertion. Snippets can then expand matching phrases.

This is a feature-availability and product-copy issue: the app describes the preset as “Raw transcription with no AI formatting,” while the live homepage shows a Verbatim control. The observed 1.0.0 (12) Formatting screen does not show it. A user needing a literal quotation or evidentiary transcript should compare versions rather than trust a label or website mockup.

A two-minute check

  1. In the signed app, first check whether a Voice to text or Verbatim control is actually present. Record its exact label and app version. If it is absent, do not assume the preset is available; test the default local-cleanup path instead. Use a fictional local note and document snippets that might match the test phrase.
  2. Dictate: “Um, the client did not approve twelve percent. I said twelve, not twenty.”
  3. Inspect the inserted result and the history versions where available. Mark any change to “did not,” the two numbers, or the speaker's phrasing.
  4. If the signed app exposes a second writing mode, repeat with the same model and microphone. Record the active formatter and endpoint. Do not compare a local-cleanup result to a cloud-polished result as if only the speech model changed.

For legal quotes, audit walkthroughs, adviser instructions, or incident notes, preserve the speaker's meaning and follow the organization's source-record procedure. A cleaned note may be useful for reading; it is not automatically a verbatim record. If a literal transcript is required, use a tool and workflow that has been tested for that requirement, and retain the original recording only where policy permits it.

What the mode does and does not control

StageVoice to text path
Speech recognitionWhisperKit model on the Mac, using selected model and language.
Optional formatterMode selection forces the formatter to None; it does not call the chosen Cloud API or Ollama formatter for that preset.
Local cleanupStill runs when the formatting pipeline flag is enabled.
SnippetsCan still expand matched trigger text later in the pipeline.
Insertion and historyFinal text can enter clipboard-backed paste and local history; destination rules still apply.

This table describes what the preset would do if the signed app exposes and activates it. It is not a guarantee of current feature availability or remotely configured feature flags. The Default provider is the visible local-cleanup choice in the 1.0.0 Formatting settings. See local versus cloud formatting for provider and endpoint checks.

Questions people ask

Does Voice to text send my transcript to a hosted AI formatter?

If activated, the preset forces the formatter to None. Its availability in the signed app is unverified. Account, telemetry, update, model-download, and destination-app traffic remain separate. Verify the installed control and configuration for confidential work.

Why did it remove an “um” if the mode says raw?

The app still runs a deterministic cleanup pass when no provider is selected, so the result is not untouched recognition output; the history view can help compare the raw recognition text with the inserted text.

Is Voice to text suitable for a legal quotation?

Only after a release-specific test and human review. Punctuation, fillers, names, and numbers can differ from speech. A polished output should not silently replace an evidentiary source.

Sources

Keep going

Read next

More how to
Put the guide to work

Make the first useful sentence.

Set up SuprFlow on your Mac, then try this guide in the app where you actually write.

  1. 01Download a speech model
  2. 02Dictate a harmless sample
  3. 03Check the text in its destination
SPECIAL OFFER$29 once Original price: $125

One license. Up to 5 Macs you own.Special offer, down from $125.Apple Silicon · macOS 14+

Get SuprFlow for Mac

7-day no-questions-asked refund on your first purchase. Contact support within seven days for a full refund. Read the refund terms ↗