Wispr Flow Release Notes
51 release notes curated from 54 sources by the Releasebot Team. Last updated: Jul 15, 2026
- Jul 9, 2026
- Date parsed from source:Jul 9, 2026
- First seen by Releasebot:Jul 15, 2026
Reliability and accuracy: where things stand
Wispr Flow improves dictation reliability with 99.9% uptime, 30% lower latency, and a fix for overly aggressive Auto Cleanup accuracy issues. It also lets users move the Flow Bar to the left or right edge and adds a Canadian English nudge with one-tap dictionary switching.
Reliability and accuracy are the core of what Flow does. Here's where things stand:
- Uptime: Dictation stayed up and running 99.9% of the time over the past few weeks.
- Speed: Dictation latency is down 30% since the start of the year, and it's still coming down.
- Accuracy: We tracked the biggest driver of recent accuracy issues to an Auto Cleanup setting that was too aggressive for some users. That's fixed now.
Put the Flow Bar wherever works best for you
The Flow Bar used to sit fixed at the bottom of your screen, which meant it could land right on top of something you needed next, like the send button in Gmail. Now you can drag it to the left or right edge instead, so it's out of the way of whatever you're actually doing. On Mac, that also means it's no longer sitting on top of your dock.
It remembers your position too, so it stays put instead of resetting every time you reopen Flow.
Just click and drag the Flow Bar to reposition it.
Canadian English
English-speaking users in Canada now get a friendly, one-time nudge to switch to Canadian English. A single tap swaps in the Canadian dictionary so spellings like "colour" and "centre" come out just right.
We shipped this on Canada Day. Seemed like the right day for it.
Find it: Go to Settings > General > Dictation Languages.
Original source - July 2026
- No date parsed from source.
- First seen by Releasebot:Jul 3, 2026
Flow now switches back automatically for more apps
Wispr Flow adds Auto-switchback to return users to their host app after dictation across iOS versions and more apps.
Auto-switchback automatically returns you to your host app the moment you finish dictating, so you're not stuck manually tapping back in. It now works across every supported iOS version, including the iOS 27 beta.
We've also added native switchback support for a long list of apps that didn't have it before, including Claude, ChatGPT, Gemini, Grok, Perplexity, Kin, LinkedIn, and many more messaging and productivity apps.
Original source All of your release notes in one feed
Join Releasebot and get updates from Wispr Flow and hundreds of other software products.
- July 2026
- No date parsed from source.
- First seen by Releasebot:Jul 3, 2026
Let Flow pick the right microphone automatically
Wispr Flow now supports ranked microphones and seamless fallback when headsets are plugged in or unplugged.
Rank your microphones once and Flow uses your top available one without you switching manually. Plug in a headset and Flow picks it up; unplug it mid-dictation and Flow falls back to your next-ranked mic without interrupting you.
Find it: Settings > General > Microphone
Original source - July 2026
- No date parsed from source.
- First seen by Releasebot:Jul 3, 2026
Keep dictating when you dock your laptop (clamshell mode)
Wispr Flow now auto-switches to your top external mic when you close your laptop lid and back to built-in when you reopen it.
Close your laptop lid and Flow switches to your highest-ranked external mic on its own instead of interrupting you with a notification, then returns to your built-in mic when you open it again. Reordering mics in settings now applies instantly, with no need to restart recording.
Find it: Settings > General > Microphone
Original source - July 2026
- No date parsed from source.
- First seen by Releasebot:Jul 3, 2026
Never Lose a Dictation Again
Wispr Flow adds clearer dictation recovery with notifications, sound alerts, and one-tap retry for failed transcripts. It also brings an always-on Live Dictation timer for all dictations and smarter, faster backspace behavior across most languages.
Interrupted dictations and retries
If you've ever started a dictation outside the keyboard and had a phone call, Siri, or the five-minute limit cut you off, you know the worst part was the silence: you had no idea whether your words were saved. Now Flow tells you. When something interrupts a dictation, Flow sends a notification with sound so you know your transcript made it to your history. Tap it and you'll jump straight to that transcript, already copied to your clipboard, with a one-tap copy button right on the row. If a dictation fails instead of finishing, you get a "Tap to retry" notification that kicks off an instant retry.
- Interrupted? You'll know: A gentle notification with sound confirms your transcript is saved whenever a call, Siri, or the time limit cuts a dictation short.
- One tap back to your words: Tapping the notification opens the transcript, auto-copies it, and highlights a copy button on the row.
- Failures retry themselves: A failed dictation sends a "Tap to retry" notification that triggers the retry for you. (Make sure Flow has notification permission so these can reach you.)
One-Tap Retry for Failed Transcripts
Failed transcripts in your history used to be easy to miss, and recovering one meant digging through a context menu. Now they're impossible to overlook: any failed transcript shows up in orange with a "Retry your X:XX transcript" label and a dedicated retry button. Tap the row or the button to retry it instantly. No more wondering whether a transcript was empty or guessing how to get it back.
Help Center: Retry failed transcriptions
Always-On Live Dictation Timer
The Dynamic Island and Lock Screen Live Activity now show a running timer for every dictation, not just the ones you start with the Action Button. Whether you kick things off from the keyboard or anywhere else, you'll always have a clear visual signal that Flow is listening, so you can speak with confidence without peeking at the screen or restarting unnecessarily.
Smarter, Faster Backspace
The keyboard's backspace now behaves like the iOS system keyboard. A single tap deletes one character on release, and holding it down speeds up from character-by-character to whole-word deletion after about two seconds, with natural pauses at punctuation. It works across most languages (including Chinese, Japanese, Korean, Hindi, Bengali, and more), so cleaning up a long passage no longer means grinding through one character at a time.
Original source Similar to Wispr Flow with recent updates:
- Perplexity release notes27 release notes · Latest Jul 13, 2026
- OpenAI release notes845 release notes · Latest Jul 18, 2026
- n8n release notes60 release notes · Latest Jul 14, 2026
- Cursor release notes114 release notes · Latest Jul 17, 2026
- Obsidian release notes92 release notes · Latest Jul 14, 2026
- GitHub release notes654 release notes · Latest Jul 21, 2026
- July 2026
- No date parsed from source.
- First seen by Releasebot:Jul 3, 2026
Less bubble, more screen
Wispr Flow adds a smarter, more discreet bubble experience with auto-shrinking, customizable compact styles, and automatic minimization in search fields. It also brings a cleaner notification tray, a new copy button after dictation, adjustable bubble opacity, and reliability improvements for Android 16.
The Flow bubble now auto-shrinks after 5 seconds of inactivity.
Choose between a compact icon or a tiny dot, and it even auto-minimizes in search fields so it's never in the way. Tap it anytime to restore and start dictating. Find the setting under Settings 14 Bubble Size.
Cleaner notification tray
We removed the persistent foreground notification that was cluttering your notifications. Flow still runs reliably in the background 14 you just won't see the extra notification anymore.
Learn more.
Copy button after dictation
A quick copy-to-clipboard button now appears next to the bubble after each dictation. Tap it to grab your transcribed text, or let it auto-dismiss after 10 seconds.
Learn more 14
Adjustable bubble opacity
If the bubble feels visually distracting over certain apps, you can now dial it back. A new opacity slider in Settings lets you set it anywhere from 20% to 100%.
Reliability improvements
The Flow bubble is more responsive when you tap into text fields, the service recovers automatically if your device's battery manager stops it, and we're now fully compatible with Android 16.
Original source - July 2026
- No date parsed from source.
- First seen by Releasebot:Jul 3, 2026
Scratchpad syncs with Desktop.
Wispr Flow now syncs Scratchpad notes between Desktop and iPhone, with iPhone gaining a plain-text Notes editor.
Your Scratchpad notes now sync between Desktop and iPhone.
On iPhone, however, it's a plain-text editor and is still named Notes.
It will be renamed Scratchpad and include more functionality in future releases.
Original source - July 2026
- No date parsed from source.
- First seen by Releasebot:Jul 3, 2026
Less bubble, more screen
Wispr Flow adds a smarter, less intrusive bubble with auto-shrink, compact display options, and auto-minimize in search fields. It also brings a cleaner notification tray, a new copy button after dictation, adjustable opacity, and reliability improvements.
The Flow bubble now auto-shrinks after 5 seconds of inactivity
Choose between a compact icon or a tiny dot, and it even auto-minimizes in search fields so it's never in the way. Tap it anytime to restore and start dictating. Find the setting under Settings 14 Bubble Size.
Cleaner notification tray
We removed the persistent foreground notification that was cluttering your notifications. Flow still runs reliably in the background 14 you just won't see the extra notification anymore.
Learn more.
Copy button after dictation
A quick copy-to-clipboard button now appears next to the bubble after each dictation. Tap it to grab your transcribed text, or let it auto-dismiss after 10 seconds.
Learn more 14
Adjustable bubble opacity
If the bubble feels visually distracting over certain apps, you can now dial it back. A new opacity slider in Settings lets you set it anywhere from 20% to 100%.
Reliability improvements
The Flow bubble is more responsive when you tap into text fields, the service recovers automatically if your device's battery manager stops it, and we're now fully compatible with Android 16.
Original source - July 2026
- No date parsed from source.
- First seen by Releasebot:Jul 3, 2026
Quick Dictation to Notes shortcut
Wispr Flow adds a built-in Quick Dictation to Notes Action Button shortcut for direct saves to Flow Notes on iPhone.
You can now assign a built-in "Quick Dictation to Notes" shortcut to your Action Button, no iCloud workaround needed. Press the button, dictate, and your words are saved directly to Flow Notes, even when you stop from the Dynamic Island. Find it in Settings 14 Action Button 14 search for Wispr Flow.
Help Center
Set up Action Button for Flow on iPhone
Original source - July 2026
- No date parsed from source.
- First seen by Releasebot:Jul 3, 2026
Adapting to iOS 26.4
Wispr Flow adds default style controls and a Quick Style Switcher to help users adapt to iOS 26.4 keyboard changes. The update keeps writing style consistent across apps and lets people temporarily switch tone from the keyboard.
Apple changed how third-party keyboards work in iOS 26.4, and it affected two core parts of the Flow experience. We moved fast to minimize the friction and give you new ways to control your writing style.
What changed (these are Apple platform changes, not something we control):
Quick swipe to start: In many apps, tapping "Start Flow" used to briefly open the Flow app and automatically return you to what you were doing. After iOS 26.4, you'll need to swipe back to your app manually. Voice-to-text still works exactly as before once you're back in your app.
Styles no longer auto-adjust by app: Flow adjusted formatting depending on which app you were typing in (more formal in email, more casual in messaging apps). iOS 26.4 removed that ability. Your default style now applies everywhere.
What we shipped to help:
Default style settings in the app: Set your preferred writing style in the Style tab, and it applies across all apps. You can update it anytime.
Quick Style Switcher on the keyboard: A new style pill appears above the keyboard, letting you tap to temporarily switch your style for any conversation without leaving the app you're in. The override resets back to the default style after 15 minutes of inactivity.
Both features are available for users with English, British English, or auto-detect language settings.
Desktop and Android are unaffected. You'll continue to get the full seamless experience there.
[Help Center: Adapting to iOS 26.4]
Original source - Jul 2, 2026
- Date parsed from source:Jul 2, 2026
- First seen by Releasebot:Jul 3, 2026
Reliability & Stability
Wispr Flow improves Android reliability with self-healing Flow Bubble recovery, accessibility permission recovery, crash fixes, per-account isolation, cleaner bubble UX, and fixes for Xiaomi and OnePlus battery optimization and overlay issues.
We heard from Android users that things haven't felt as solid as they should, so the team spent the past few months tightening things up:
- Flow Bubble recovery: If the bubble disappears, it now recovers on its own, and you'll get a system notification if it needs a nudge.
- Accessibility permission recovery: If a device revokes Flow's accessibility permission in the background, Flow now catches that and recovers instead of silently breaking.
- Crash fixes: Fixed several crashes.
- Multi-account isolation: If you share a device with other accounts, your Flow settings and onboarding now stay isolated per account instead of carrying over between users.
- Flow Bubble UX cleanup: Fixed sizing glitches, visibility issues, errors that used to stick around, and snooze/shrink not working right.
- Xiaomi/OnePlus fixes: Resolved battery optimization and overlay problems specific to these devices.
- Jun 24, 2026
- Date parsed from source:Jun 24, 2026
- First seen by Releasebot:Jul 3, 2026
More consistent and accurate formatting
Wispr Flow fixes dictation formatting issues and now uses one current model for more consistent transcripts. Users should see cleaner course correction, fewer unwanted word swaps, and punctuation and edits that follow instructions more reliably.
Over the past few months, some dictations had formatting issues: course corrections that didn't land, words swapped for ones you didn't say, punctuation or edits getting ignored, and quality that varied session to session. A portion of dictations were being handled by older models, one too conservative on course correction and one too aggressive about changing words.
Both are now fixed. Everyone is on the same current model, so quality is consistent every time you speak. You'll notice:
- Cleaner course correction
- Fewer changes to words you said
- Punctuation and corrections that do what you tell them to
To keep it from recurring, we've added A/B testing and expanded benchmarks so every new model is tested against real usage before it ships. If you hit a formatting issue, report the transcript from inside the app with a note on what went wrong. Those reports are what help the team improve Flow.
Original source - June 2026
- No date parsed from source.
- First seen by Releasebot:Jun 24, 2026
Flow in every application
Wispr Flow brings voice dictation to every app, helping users write emails, docs, code and messages faster with AI commands, auto-edits and tone matching. It works across Mac, Windows, iPhone and Android and offers a 14-day free trial.
Work at the speed you think in every app you use. Email, messages, docs or code—Flow works in any text box.
Reply to emails and comments faster by speaking instead of typing.
Keep your outreach personal with Flow’s tone-matching feature.
Maintain focus by speaking instead of typing, reducing mental fatigue.
Capture ideas and tasks instantly before they slip away."It's like having a cheat code for clearing out your inbox. The accuracy and responsiveness feel like magic."
Conor Swanson, Code TalentUse Flow in any text box in any app
Essential apps
- Arc
Dictate seamlessly in any text box in Arc Browser for faster, hands-free workflow. - Cursor
Refine prompts, craft ideas, and streamline your workflows in Cursor with dictation. - Chrome
Get smooth and accurate dictation in every app you use in Chrome. - LinkedIn
Respond to comments and send messages on LinkedIn without typing. - Perplexity
Search, ask questions, and get concise answers on Perplexity without typing. - Notion
Capture ideas and organize projects as soon as they come to mind. - X
Share your thoughts instantly and engage with replies by using your voice. - WhatsApp
Have natural conversations that keep your authentic tone. - Apple Mail
Speak your message and get a perfectly written, formatted email—ready to send. - VS Code
Write code, navigate files, and streamline projects in VS Code with dictation. - iMessage
Dictate and send messages in iMessage for faster, more natural, error-free conversations. - Gmail
Dictate and send emails in Gmail that sound like you and formats itself. - Slack
Send quick updates to your team without breaking your focus. - Claude
Engage with Claude effortlessly by dictating prompts and refining responses for faster, natural AI interactions. - Superhuman
Get through your Superhuman inbox even faster by dictating emails with natural speech. - ChatGPT
Get the output you want by speaking your prompts directly into ChatGPT.
Documentation
- Apple Notes
Capture quick thoughts, jot down ideas, or draft longer notes with accurate dictation. - Evernote
Instantly record notes and ideas in Evernote by using your voice. - Google Docs
Accurately dictate your thoughts into Google Docs for faster, clearer writing. - Obsidian
Connect the dots faster by using your voice in Obsidian. - Notion
Capture ideas and organize projects as soon as they come to mind.
Engineering
- Cursor
Refine prompts, craft ideas, and streamline your workflows in Cursor with dictation. - Discord
Send messages, manage servers, and engage with communities on Discord using voice. - Github
Dictate commit messages, issues, and code reviews with Flow, streamlining your GitHub workflow. - Jira
Create tickets, update sprints, and add comments in Jira using voice commands for faster workflows. - Linear
Create and manage Linear tasks using your voice to stay on track with what matters. - Stack Overflow
Answer questions, draft posts, and engage with the community on Stack Overflow using natural voice input. - Warp
Execute terminal commands and scripts using your voice to make CLI feel natural. - VS Code
Write code, navigate files, and streamline projects in VS Code with dictation. - Canva
Dictate your Canva text fields with natural speech for faster, polished designs. - Replit
Build, run, and debug code collaboratively in Replit using dictation for faster development across languages. - Lovable
Design lovable developer experiences by dictating better docs, CLIs, and onboarding flows—faster and more naturally. - v0
Generate production-ready UIs in v0 by dictating component ideas and layout instructions—no manual coding needed. - Bolt
Spin up internal tools and dashboards in Bolt using dictation to describe workflows, fields, and logic in plain English.
Project management
- Click Up
Move through projects faster by dictating your tasks, docs, and comments. - Jira
Create tickets, update sprints, and add comments in Jira using voice commands for faster workflows. - Linear
Create and manage Linear tasks using your voice to stay on track with what matters. - Monday
Spend less time typing out tasks, descriptions, and comments. Speed up your workflow with voice. - Notion
Capture ideas and organize projects as soon as they come to mind. - Trello
Get things done faster by dictating your descriptions and comments within Trello.
Writing
- Evernote
Instantly record notes and ideas in Evernote by using your voice. - Google Docs
Accurately dictate your thoughts into Google Docs for faster, clearer writing. - LinkedIn
Respond to comments and send messages on LinkedIn without typing. - Medium
Create more authentic stories by speaking your first draft. - Obsidian
Connect the dots faster by using your voice in Obsidian. - Notion
Capture ideas and organize projects as soon as they come to mind. - Substack
Capture your authentic ideas by speaking your first draft in Substack. - Claude
Engage with Claude effortlessly by dictating prompts and refining responses for faster, natural AI interactions. - ChatGPT
Get the output you want by speaking your prompts directly into ChatGPT.
Start flowing
Effortless voice dictation in every application:
4x faster than typing, AI commands and auto-edits.Available on Mac, Windows, iPhone, and Android.
Original source
Free for 14 days. - June 2026
- No date parsed from source.
- First seen by Releasebot:Jun 24, 2026
Flow for Students
Wispr Flow introduces a fast, accurate voice dictation sidekick for any app, with seamless speech-to-text on MacOS, Windows, iOS and mobile. It adds AI command mode, tone control, and support for 100+ languages for faster writing everywhere.
Introducing Flow, the fastest and most accurate voice dictation sidekick that works in any application. Work at the speed you think!
Flow through finals like the genius you are.
Seamless speech-to-text in every application on your computer and phone.
4X faster than typing
Save hours on notes and essays by using your voice to type.
Sounds like you
Set casual or formal tone in any app.
Works everywhere (+ mobile)
Flow works in Docs, Apps, and notes—no new software to learn.
Command mode with AI
Fine-tune and edit your writing with AI to sound your very best.
Get 3 months free for finals
Auto renews at 50% off. Cancel anytime.
How to sign up
- Sign Up
Click “Download for Windows” and sign in to create an account.
- Download the App
Choose where you want to Flow: now available on MacOS, Windows, and iOS.
- Get Verified
Sign in to the app with your edu email—we’ll auto-activate your student offer: 3 months free, then $7.2 /month (50% off). Cancel anytime.
3-month trial is for new subscribers only
Flow Pro for Students
$7.2 /mo billed annually
50% off, just for students
Use all of our features on our Pro Plan
Turn thoughts into action
Flow is perfect for quickly capturing raw ideas and getting started outlining papers. Just speak your thoughts and Flow will organize them, ensuring nothing is missed.
Write with precision
Be the best version of yourself
Flow understands that academic writing often involves complex terminology and precise language. Whether you're crafting essays, research papers, or presentations, Flow’s context-aware dictation ensures accuracy with tricky words and specialized terms.
Say goodbye to the frustration of typos and hello to polished writing. Flow will help you sound like the best version of you!
AI that helps you get that A+
Flow’s Command Mode supports editing and formatting of your content so it’s fit for purpose:
- Shorten, lengthen, or edit your writing to take something from good to great.
- Provide more context when writing prompts for tools like ChatGPT to get more relevant and refined results.
- Ask Flow for more context about anything.
Effortless, never awkward
Group projects? Randomized teams? That weird panic when you have to message someone you’ve never talked to before?
Flow helps you get the words out before overthinking kicks in. Whether you’re:
- Breaking the ice in a new group;
- Messaging a professor without rewriting it over and again; or
- Sending updates without staring at a blinking cursor
Talk it out. Ship it faster.
Flow is your voice-to-dev pipeline—speak your logic, and it writes clean, structured text for code, docs, or AI prompts.
Whether you're:
- Spitballing algorithm logic before opening your IDE
- Penning down a quick function
- Drafting prompts for Cursor, Copilot, or ChatGPT
- Or just rubber ducking your way through a gnarly problem
We took Flow to campus
Love letters to Flow
Flow’s basically my brain dumb button. I dump all my chaotic thoughts—half ideas, random 2am brain sparks—and it just keeps up. Later, I go back and realize some of it’s actually usable. Like, real essay material. Who knew. — Aryaman Sharma, NYU Courant
What makes it really magical is that it's just so natural. With any products that I try, I have to think about how I can get it to do what I wanted to do. But with flow there's none of that. Whatever I think- about what I want to write is whatever ends up getting written, which is an unbelievably cool experience. — Yash Mishra, Student at University of Washington
I used to second-guess every sentence when writing to someone new—especially in work settings. Flow fixes that. It’s like having someone gently tap you on the shoulder and say, “Hey, you’re good. Hit send.” — Ibrahim Khursheed, University of Sydney
I’ve been using Flow for a while now and absolutely love it. Whether it’s for emails, ChatGPT, or coding, it has doubled my productivity and speed. I can even use it within my favorite IDE to vibe code!! Thanks, Flow! — Ahmet Dedeler, High School Student
Frequently asked questions
Do I qualify for a student discount even if I don’t have a .edu email?
Yes! We accept alternate forms of student verification like a student ID, enrollment confirmation, fee receipt, or a screenshot of your student dashboard. If you're studying at a university without .edu emails or are attending online programs, you're still eligible.
Fill out the form here: Enable the Education Discount for your Flow account
What languages does Flow support? Can I use it if I write in multiple languages?
Yes! Flow supports 100+ languages for dictation. A lot of students use it to write in multiple languages—like English and Spanish or French and Arabic. You can switch between them easily, and Flow auto-detects as you go.
What do I get when I refer a friend?
You and your friend both get a free month of Flow Pro when they hit 2,000 words or upgrade to Pro. Just share your referral link with your friend — you can find it in your desktop app under “Refer a Friend” on the home screen.
Can I use headphones with Flow?
Yes, you can! But for the best results, we recommend using your device’s built-in mic or a wired one instead of a bluetooth connection.
Read more here.
What can I run Flow on—mobile, desktop, both?
Flow works on MacOS, Windows and iOS for now!
Is Flow accessible for students with disabilities?
Absolutely. Flow is designed with accessibility in mind and supports a wide range of needs and learning styles. Many students who face barriers to traditional input methods have found Flow to be a powerful tool for expressing their thoughts.
Can educators or non-profit staff also apply for a discount?
Yes! If you're a teacher, researcher, or work with a non-profit, you can apply for our educator/non-profit discount using your .edu email.
If you don’t have one, just fill this form to get access.
Can I use Flow to transcribe audio or video files like lecture recordings?
Not right now—we don’t support file uploads for transcription just yet.
I’m on a tight budget. Is there any way to get Flow for free or at a lower cost?
We offer a sweet student discount, and in special cases (like accessibility needs or financial hardship), we're always happy to work with you. One of the easiest ways to get Flow for free is by referring a friend — once you both hit 2000 words, you'll both get a free month!
Interested in becoming an Ambassador?
We're cooking up a new program that takes Flow to your campus. Share your email with us, and we'll be in touch over the summer.
Apply now
Original source - Jun 17, 2026
- Date parsed from source:Jun 17, 2026
- First seen by Releasebot:Jul 3, 2026
A more reliable Wispr Flow
Wispr Flow improves stability and accuracy with infrastructure scaling, better language routing, cleaner audio handling, and stronger model evaluations. It also introduces Cloud Sync, a new data control that enables cross-device sync and future personalized features while keeping Privacy Mode separate.
Rapid growth in our user base strained our infrastructure recently, and some users saw slower dictation, lower accuracy, or trouble signing in and using polish.
We've scaled up our infrastructure to keep pace with growth and isolated the most demanding work onto a dedicated system, so heavy usage in one area can't disrupt the rest. We've also tightened monitoring to catch and resolve issues faster. This work is ongoing.
On accuracy, we fixed several specific issues:
- Better language routing: UK English and Swiss German weren't always being sent to the right place, which hurt accuracy for those users. Both are now routed correctly.
- Cleaner audio handling: Audio compression wasn't fully working as intended, which could affect how well your speech was transcribed. It's now working the way it should.
- A higher bar for every model: We built a much stronger set of evaluations across all of our models, so we can catch accuracy regressions before they reach you.
Some of the recent drop in accuracy was a side effect of this infrastructure strain rather than the models themselves, and it improves as the stability work lands.
We're also testing a new round of model improvements to push accuracy further, and building toward personalized speech models that adapt to how you speak over time.
Updated data controls
We're introducing Cloud Sync, a new setting alongside Privacy Mode that gives you more granular control over how Flow handles your data.
Why we're making this change: To support upcoming features like personalized speech models, Flow needs to store your transcription data on our servers. Until now, Privacy Mode meant zero data retention: nothing stored, nothing trained on. That was the right default when Flow was purely a dictation tool, but it also meant we couldn't build features that depend on your history, like syncing notes across devices or learning how you speak to improve accuracy over time.
Rather than weaken Privacy Mode, we split it into two independent controls:
- Privacy Mode controls whether your data can be used for model training. This hasn't changed. With Privacy Mode on, none of your data is used to train or improve AI models, by us or any third party.
- Cloud Sync controls whether your transcription data (transcripts, audio, and dictation history) is stored on Wispr's servers. Turning Cloud Sync on unlocks cross-device sync for Wispr Scratchpad and enables future features.
What this means for existing users:
- If you had Privacy Mode on, nothing changes automatically. Privacy Mode remains on, and Cloud Sync stays off. This is identical to the previous level of privacy 14 zero data retention for all audio and transcript data.
- If you had Privacy Mode off: Nothing changes. Your settings carry forward as-is.
- When you're ready to try features that require Cloud Sync, or want to enable cross-device sync, Wispr will prompt you to turn it on. You're always in control.
Your dictionary, snippets, and custom prompts continue to sync across devices, regardless of your Cloud Sync setting. Wispr Scratchpad requires Cloud Sync to be enabled for cross-device sync.
For the full breakdown, see our updated Data Controls page.
Original source
Curated by the Releasebot team
Releasebot is an aggregator of official release notes from hundreds of software vendors and thousands of sources.
Our editorial process involves the manual review and audit of release notes procured with the help of automated systems.