Transcription suddenly got worse or feels less accurate
Last updated: November 2, 2026
Available on: Mac, Windows, iOS, Android
If dictation loses words, produces garbled text, or suddenly feels less accurate, compare a short recording with its transcript. This separates a microphone problem from a language or transcription problem.
Quick checks
Restart Flow: quit and reopen it, then dictate a sentence longer than three seconds; shorter recordings can return empty text.
Check the active microphone: a connected headset, webcam, or Bluetooth device can change the input between dictations.
Record in another app: use Voice Memos on Mac/iPhone or Sound Recorder on Windows. If audio is unclear there too, fix the microphone or system input first; if it is clear, continue below.
Check desktop audio
Check | Mac | Windows |
Microphone permission | System Settings → Privacy & Security → Microphone | Settings → Privacy & security → Microphone; allow desktop app access |
Input device and level | System Settings → Sound → Input | Sound control panel → Recording |
Test the same phrase after each change. If the text is accurate, stop; otherwise continue.
Enable Flow's microphone access using the permission settings above, then reopen Flow.
Select your intended microphone in the system sound settings.
Set the input level so the meter moves clearly without constantly reaching maximum.
Open Flow's Settings → General → Microphone → Change.
Select the intended mic and speak; the selected row's live level meter should move. If it barely moves, raise the input level or move closer to the mic.
Compare a built-in or wired mic against Bluetooth if accuracy remains poor.
Flow does not adjust the system input level, room noise, or microphone distance for you. Built-in or wired microphones are recommended over Bluetooth: AirPods can add delay and hurt accuracy.
Note: Bluetooth microphones can delay audio capture and clip opening words — on Mac, let the mic become ready before speaking. For clipped beginnings, see Missing first words in transcriptions.
With a MacBook lid closed, the built-in mic is unavailable and cannot be selected. Flow picks an available external mic; otherwise choose Select microphone on “Your laptop lid is closed”.
If Flow reports no audio or cannot open the mic
On Mac, Grant Permission requests access, or opens privacy settings after a denial. On Windows, Open Settings opens microphone privacy settings.
No microphone detected: connect a mic or choose Open Sound Settings to check the input, not privacy permissions.
Selected microphone is unavailable: use Choose Microphone to select a connected device.
Microphone unavailable / Microphone error: close other mic-using apps or reconnect the mic, then retry; Troubleshoot opens microphone help.
Unable to access mic: after more than eight seconds, Restart App is offered; the prompt clears if access succeeds.
Is your microphone muted?: shown when a recording captures nothing; it names the mic and offers Select microphone and Troubleshoot.
After changing the input or access settings, dictate the test phrase again; if audio remains absent, contact support below.
Listen to the original and retry
On desktop, open Home and use Search past transcripts. Hover over or keyboard-select an entry to reveal its actions.
Important: Desktop audio is cleared after 14 days; transcript text remains. Save any audio you need before then — after 14 days, playback is hidden and Extract audio is disabled. Retry replaces the entry's text. If your or your organization's policy prevents storage, “History disabled” appears and no audio is saved for playback, extraction or retry.
Select Play audio to compare the recording with its text.
Choose More options (⋯) → Retry transcript if the audio is complete and clear.
Compare the updated text; if it remains wrong, check languages below.
If the recording has the same gaps, check mic position, Bluetooth, input volume and background noise. More options (⋯) → Extract audio saves a WAV file.
When retry is unavailable or fails
Every retry needs retained audio: missing audio prevents playback, download and retry, and recordings dismissed under about half a second are never saved.
The menu's Retry transcript has no minimum duration but is disabled while the row loads or dictation is active. Inline Retry on empty or failed entries needs at least five seconds of audio and an entry newer than 14 days; Recover for dismissed or still-processing entries has that age limit but no minimum duration.
For “Cannot retry while dictating”, stop dictation first. “Retry failed. Please try again.” is a generic failure. A retry that times out or is taken over by a new dictation leaves the original text unchanged.
If a long recording is incomplete
Desktop dictations stop at 20 minutes, with a warning at 19 minutes; the limit-reached notification offers recovery of the last text.
Important: Very long recordings can lose final words after you stop. Retry when available, or divide long thoughts into shorter recordings.
If text exists but never reached the app, follow text-pasting troubleshooting instead.
Check phone audio
iPhone
After microphone access is denied, only iOS Settings can restore it; the record button stays disabled until access is granted.
Tap Go to Settings on “We can't hear you”.
Enable Microphone in iOS Settings.
Return to Flow; permission is rechecked without restarting the app.
Test with Bluetooth disconnected, the phone's microphones uncovered and less background noise.
Swipe a History entry to use Retry, then compare the updated text.
Retry is disabled while already running. A failed entry can also be tapped to retry.
If audio was never recorded, it cannot be recovered — dictate again.
Android
Confirm your intended microphone or hearing aid is connected before recording. If accuracy drops partway through, the input may have switched: stop, confirm the active input, and start again. When Android will no longer show its permission dialog, Flow displays “We can't hear you”.
Tap Go to Settings to open Wispr Flow's system App info page.
Allow microphone access there.
Test another phrase and compare the result.
Choose ⋮ → Retry Transcript on a failed entry with captured audio, then check the updated text.
While a retry runs the entry shows “Retrying…” and the Retry option is hidden, so the same recording can't be retried twice. If the retry finds no speech, the entry is saved as an empty transcription. Retrying from the error card on the dictation bubble makes one fresh attempt; pressing to dictate meanwhile shows “Still retrying your last dictation. Try again in a moment.”
Without captured audio, Retry Transcript is hidden. “Retry failed. Audio file not found.” or “Retry failed. Audio file missing.” means the recording is unavailable; for “Retry failed. Please try again.”, retry.
Check language and repeated substitutions
Mac and Windows languages
Keep at least one language or Auto-detect selected. Auto-detect selects no individual languages; turning it off during the same dialog session restores your previous choices.
Open Settings → General → Languages, also labelled My languages.
Turn off Auto-detect to search and select individual languages.
Select the languages you speak.
Choose Save and close.
A conflicting variant replaces the previous selection: English - British replaces another English variant.
If Ukrainian speech comes back as Russian, select both Ukrainian and Russian. Selecting Ukrainian prevents language-mismatch notices; their coverage does not include every available language.
iPhone languages
Open Flow's Settings → Set language and select the languages you speak in the searchable list. Selections apply immediately; there is no Auto-detect toggle or save step.
Android languages
At least one language is required, and there is no Auto-detect toggle.
Open Flow's Settings → Languages.
Tap Add more.
Select the languages you speak using Search languages.
Tap Done, then Save — changes apply only after Save.
Dictate the same phrase again to check the result. See language setup for more.
For repeatedly incorrect names or technical terms, review your dictionary words and corrections. An empty replacement removes the matching word; empty or whitespace-only phrases are ignored. Separate capitalization entries are unnecessary.
Still stuck?
Contact support if clear, complete audio repeatedly produces incorrect text after these checks, microphone access still fails, or retry or report submission fails.
Open Help → Talk to support on desktop.
Select Report an issue; Request Type → Issue is preselected.
Describe your platform, microphone, languages, affected recording, expected text, actual result and checks completed.
Choose Send message.
A message is required; you can attach one image up to 10 MB. If sending fails, Flow copies your message to the clipboard: paste it and try again.
Tip: To flag a bad transcript without requesting a reply, select the desktop entry's flag icon, describe what you expected instead, and choose Send feedback. An empty report cannot be sent, and this feedback does not receive a reply.