Voice Profile updates: when the 'Update your Voice Profile' notification appears

Available on: Mac, Windows

Voice Profile is an Insights report built from your dictation. If the update prompt hasn't appeared, or the remaining word count has gone up instead of down, check the milestone and stored-history requirements below before generating a new report.

When the update prompt appears

The "Update your Voice Profile" notification is for an existing profile; first-time users see "Discover your Voice Profile". Eligibility requires both a new cumulative word milestone (the first is 2,000 words; later milestones are higher) and enough completed dictation still stored on this device:

Cumulative milestone	Minimum locally stored words
Up to 2,000	2,000
Above 2,000 through 12,000	5,000
Above 12,000 through 37,000	10,000
Above 37,000	20,000

All locally stored completed transcripts with text count, including archived dictations, with no recency window: raw transcripts saved with AI formatting turned off, plus fallback, browser-extension and command (instruct) dictations. Pending, empty, dismissed and errored dictations don't. Removal of audio after 14 days doesn't remove transcripts or reduce this total.

If stored history is short, the prompt is held and the remaining count reflects the shortfall.

The prompt appears shortly after the dictation that meets the requirements — no restart or settings visit needed — and appears once per milestone.

Milestones in Settings → System is on by default. Turning it off silences Voice Profile notifications without hiding the hub card or Your Voice tab.

Note: An organization's local-data deletion policy makes Voice Profile unavailable and hides the Your Voice tab and hub card. Counting needs normal local dictation storage, so policies that delete data after 24 hours or never store it stop progress entirely.

What the progress bar counts

Only desktop dictation contributes to the report or progress; mobile dictation doesn't count. Each counted dictation lowers the remaining number, and a long dictation stops progress at the next milestone even if it could cross several.

State	Progress label
Before your first profile	"Generates in N words"
Between updates	"Next update in N more words"
Generation becomes available	"Update ready!"

After all milestones, counting stops, no generation button is offered, and the full bar has no right-hand label.

For your first profile, at least 2,000 lifetime desktop words can make you eligible on your next dictation, provided at least 2,000 words remain stored locally; otherwise counting continues normally.

On opening Your Voice, the remaining count may be blank briefly while it loads.

Warning: Deleting dictation history from your device reduces the material available for a refresh, so the counter can go back up. Flow reflects that shortfall when a dictation reaches a milestone or you try generating.

Generate an updated Voice Profile

The notification's Update profile button opens Your Voice (See profile for a first profile). You can also open the tab directly:

Open Insights → Your Voice in Flow.
Select Regenerate profile when available — Generate profile for your first profile.

Generation starts immediately if the requirements are met, and the button is disabled while building, so a second attempt won't start another report. You can keep dictating and close the window.

When it's ready, a separate "Your insights are ready!" notification offers View insights. The new profile uses your most recent locally stored dictation and replaces the one shown in Flow.

How your dictation is used

Creating a report sends selected locally stored dictations to Flow for processing, using the same dictations that count toward the milestone, including raw transcripts saved with AI formatting off. With usage-data sharing off, the card displays: "Your dictations are private and only stored locally. Never shared or stored by Wispr Flow. Your report is based on local data."

If generation doesn't start or finish

Not enough stored dictation

Generation doesn't start when the dialog says "You only have N words stored locally." and "You will need to dictate N more words so you have enough data to build your profile." The remaining count adjusts to this shortfall; this is expected.

Select OK to dismiss the dialog.
Dictate the additional words shown.
Select Generate profile or Regenerate profile again when available.

If generation starts, the tab shows "Insights are building…".

Generation failed or timed out

Generation can fail, including when no stored dictation is available. The tab shows "Something went wrong. Try again."; the hub card shows "Voice Profile unavailable" / "Keep dictating and we'll try again" with no retry button.

Open Insights → Your Voice.
Select Generate profile to retry.

If the retry starts, the tab shows "Insights are building…". A failed or timed-out update keeps your existing earlier profile.

Still stuck?

Contact support if:

The remaining word count doesn't decrease after several dictations.
Generate profile or Regenerate profile doesn't start generating and you don't see the insufficient-data dialog.

Include your platform (Mac or Windows), your device, what the progress bar shows, and the steps you tried.