Voice Profile updates: when the 'Update your Voice Profile' notification appears
Last updated: October 20, 2026
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 increased, check eligibility and stored history before generating another report.
When the update prompt appears
Existing profiles receive "Update your Voice Profile"; first-time users see "Discover your Voice Profile". Eligibility requires both a new cumulative word milestone and enough completed dictation 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 |
Milestones are 2,000, 12,000, 37,000, 87,000, 162,000, 262,000, 362,000, 462,000, 562,000 and 662,000 words. After the 662,000-word profile completes, no further prompts or updates are available.
The prompt appears once per milestone, about five seconds after the qualifying dictation; no restart or settings visit is needed. Progress toward another milestone pauses while a previous profile is generating or has failed, resuming after completion. Progress toward the first milestone is exempt from this pause.
The prompt is suppressed while the Flow window is focused on Insights → Your Voice. If another dictation or notification prevents delivery, Flow retries after the next dictation.
Milestones in Settings → System is on by default. Turning it off silences Voice Profile and other milestone notifications, including word-count, streak and referral activity, without hiding Your Voice or the Home card.
Dismissing an eligible Home card hides it for that milestone cycle; it returns at the next milestone target.
Note: Voice Profile requires normal local dictation storage. Your own or your organization's Delete after 24 hours or Never store choice hides Your Voice and the Home card and stops progress. The stricter choice applies. These choices also delete history after 24 hours or prevent it from being kept, reducing or removing the stored-word total.
What the progress bar counts
Only desktop dictation contributes to the report and progress; mobile dictation doesn't count. Totals use spoken words, not formatted or edited text. Each counted dictation reduces the remaining count, stopping at the next milestone even if it could cross several.
Completed transcripts stored with text count regardless of age, including archived dictations, raw transcripts with AI formatting off, fallback, browser-extension and command/instruct dictations. Pending, empty, dismissed and errored dictations don't count. Audio removal after 14 days doesn't remove transcripts or reduce the total.
For a first profile, at least 2,000 lifetime desktop words can make you eligible on your next dictation if at least 2,000 words remain stored locally; otherwise counting continues normally.
State | Progress label |
Before your first profile | "Generates in N words" |
Between updates | "Next update in N more words" |
Generation becomes available | "Update ready!" |
On opening Your Voice, the remaining count can stay blank for up to 15 seconds while loading. If the word count still hasn't loaded after that, Flow treats the accumulated count as zero and calculates the remaining words—for a first profile, "Generates in 2,000 words". Progress can also jump on launch to reflect a completed milestone from another session or device; an older profile won't replace a newer one.
After all milestones complete, counting stops and the full bar has no remaining-word label or generation button. A failed attempt can still show Generate profile, but selecting it does nothing if no milestone remains.
Warning: Deleting dictation history reduces the material available for updates. At a milestone or generation attempt, insufficient stored history holds the prompt and adjusts remaining words to the shortfall, so the count can increase.
Generate an updated Voice Profile
The notification's Update profile button (See profile for a first profile) opens Your Voice. To open it directly:
Open Insights → Your Voice in Flow.
Select Regenerate profile when available — Generate profile for your first profile.
Generation starts immediately when eligible, without confirmation. Another attempt is disabled while it runs. You can keep dictating and close the window.
The new profile uses your most recent locally stored dictation and replaces the one shown in Flow.
The separate "Your insights are ready!" notification offers View insights, opening Your Voice. It waits about five seconds, skips active dictation and retries later, but isn't suppressed while you're viewing Your Voice. Profiles already completed and recovered on launch or another device appear without this notification.
How your dictation is used
Creating a report sends selected locally stored dictations to Flow for processing, including raw transcripts saved with AI formatting off. Selection uses the same types of dictation that count toward eligibility, newest first; not all counted dictation is sent. Flow also receives your timezone, peak day/hour, top app at that time, most-used word and most-corrected word.
With usage-data sharing off, ready-to-generate, failed and completed-profile views display: "Your dictations are private and only stored locally. Never shared or stored by Wispr Flow. Your report is based on local data." This note isn't shown while locked or building, or when usage-data sharing is on.
If generation doesn't start or finish
Not enough stored dictation
The dialog "Not enough data to refresh your profile" can appear from Your Voice or the Home card. Generation hasn't started. It 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."
Select OK to dismiss the dialog.
Dictate the additional words shown on this desktop device.
Select Generate profile or Regenerate profile again when available.
Generation failed or timed out
Generation can fail even with enough words, including when no usable saved dictation remains or it times out after about three minutes. Your Voice shows "Something went wrong. Try again."
The Home card shows "Voice Profile unavailable" / "Keep dictating and we'll try again", without a retry button, progress bar or dismiss control. Selecting the card opens Your Voice.
Select Generate profile in Your Voice to retry, even if you already have a profile.
A failed or timed-out update doesn't erase your earlier profile.
Flow quit during generation
Within the same app session, recovery attempts are at least five minutes apart; an offline or failed connection leaves the current state unchanged.
Reopen Flow after it quits or crashes during generation.
Flow checks for completion, continues waiting, or restores your earlier profile. It shows a failure only if no earlier profile exists. If recovery times out while you have an earlier profile, that profile remains available.
"Insights are building" confirms that generation or a retry has started.
Still stuck?
Contact support if:
The remaining count stays unchanged after several dictations despite meeting the requirements above.
Generate profile or Regenerate profile doesn't start generation and no insufficient-data dialog appears, with milestones still remaining.
Include your platform (Mac or Windows), device, progress-bar text and steps tried.