Changelog
Every released version for Mac and Windows, newest first. Technical details included, because measurements belong to a change.
Jump to: VoxaFlow for Mac · VoxaFlow for Windows
VoxaFlow for Mac
1.0.0-rc8 · November 16, 2026
Fixed from feedback
- Voices in the background: when someone in the same room spoke, say a few metres away, their words sometimes ended up in the middle of your dictation. VoxaFlow now leaves out speech that is much quieter than your own voice before recognition. You can turn this off in Audio with Ignore quiet voices in the background. When two people talk at the same time, their voices cannot be separated. Thanks for the report!
1.0.0-rc7 · November 11, 2026
Overview
- With iCloud sync, the overview under All devices also shows dictations from the iPhone with their own icon: iPhone keyboard, iPhone Action button and iPhone notes
Feedback
- Send Feedback is now always at the bottom of the main window’s sidebar and stays at the top level of the menu bar menu
Fixed from feedback
- Hands-free by double tap: a very fast double tap on the dictation key was sometimes not recognized, mostly while the microphone was still starting (for example with a Bluetooth headset). Now it counts when you pressed the key, not when VoxaFlow got round to noticing. Thanks for the report!
1.0.0-rc6 · November 8, 2026
New: Notes
- The main window has a Notes section: write notes or speak into them with Dictate or your dictation key, search them and send them on with Share
- With Pro, Summarize sums a note up in a few sentences and bullet points. Apple Intelligence does this on your Mac (macOS 26 or later); nothing leaves your device. With the Medical package on, VoxaFlow does not summarize notes
- With iCloud sync you have the same notes on all your Macs. Note sync can be switched off on its own under iCloud & Data; with professional secrecy on, notes stay on the Mac
- Backups now include your notes
New: professional secrecy can be turned off
- If you need the word lists of Medical or Legal & Tax but are not bound by professional secrecy, turn off Professional secrecy under Packages or iCloud & Data. History and notes are then kept and synced via iCloud as without the package. The setting applies to this Mac only
Fixed from your feedback
- AirPods and other Bluetooth headsets: when a headset connected or disconnected, for example when your AirPods switched over to the Mac or a browser took over their microphone in a video call, VoxaFlow could freeze and only a force quit helped. Now VoxaFlow stays responsive, even in the middle of a dictation. Thanks for reporting!
1.0.0-rc5 · November 3, 2026
Fixed from a user report
- English words in the middle of a German dictation: speaking German quickly with many English technical terms, for example when dictating into Claude Code, sometimes produced half-English sentences. The speech recognition heard German words as similar-sounding English ones, such as “by” for “bei”, “so that” for “sodass” or “here any” for “hier keine”, and the AI cleanup left them in. With a “not” among them, VoxaFlow even inserted the dictation without any cleanup. The AI cleanup now brings such words back to German; technical terms, names and code stay. Thanks for the report!
AI cleanup
- If you dictate an English sentence within German, such as a commit message, a quote or a code comment, it now stays in English word for word. VoxaFlow Lektor sometimes translated such sentences into German along with the rest (“fix the login bug” became “Fixe den Login-Bug”). If the model translates anyway, VoxaFlow inserts the text with the basic cleanup only
1.0.0-rc4 · November 3, 2026
VoxaFlow Lektor
- New models for the AI cleanup: VoxaFlow Lektor replaces Qwen3 4B, VoxaFlow Lektor Kompakt replaces Qwen3.5 2B. Both are built on Qwen3, trained for dictation and, as before, run only on your Mac
- Lektor Kompakt makes far fewer mistakes than Qwen3.5 2B. For everyday dictation it is as accurate as VoxaFlow Lektor and takes about half as long; with medical and legal terms it is slightly weaker
- Your choice stays: if you had Qwen3 4B, you get VoxaFlow Lektor; if you had Qwen3.5 2B, Lektor Kompakt. Macs with 8 GB of memory still get the small model by default
- On the first start after the update, VoxaFlow downloads the new model (2.3 or 1 GB). As soon as it runs, VoxaFlow deletes the previous Qwen models, freeing up to 3.5 GB
AI cleanup
- If you correct yourself with “no, not …, but …”, the corrected version comes through: “cheese, no, not cheese, but butter” becomes “butter”. Before, VoxaFlow often inserted such dictations without the cleanup
- Numbers you say digit by digit, such as “zero one seven six, two three four …”, come through as figures. Before, VoxaFlow often inserted such dictations without the cleanup
1.0.0-rc3 · October 28, 2026
AI cleanup
- A “please” or a negation such as “not” or “never” is no longer lost in the cleanup. Before, when a sentence had two “bitte” (German “please”), the second one was often missing. If the AI still leaves one out, VoxaFlow inserts the text without its cleanup
Inserting
- Dictating into a Screen Sharing window of another Mac now inserts the text there. Before, only a “v” appeared. VoxaFlow types the text with this Mac’s keyboard layout, umlauts, “@”, “€” and accents included; the other Mac needs the same layout. Your clipboard stays untouched
- Spoken e-mail addresses are recognized more reliably, also when speech recognition breaks them up: “Holger at Login minus Online Punkt com”, “holger atlogin-online.com” or “Kanzlei Punkte” (for “Punkt de”) give a correct address. An address at the start of a sentence stays in lower case
- The last word of a dictation is lost less often, for example the “de” at the end of an address
- A short dictation in Screen Sharing no longer turns into a hands-free dictation by itself
License
- Sales start on February 1, 2027 instead of January 1. Until then Pro and all packs stay free, until January 31, 2027, also for new installations. The trial starts with the sales start
- The License page, menu, welcome window and guide now speak of the free period until January 31, 2027. Everyone who uses VoxaFlow before the sales start gets the launch price
- The License page shows your early-user code for the launch price. Everyone who uses VoxaFlow before February 1, 2027 gets the launch price on every purchase until February 28, 2027: on this Mac automatically, on another computer with the early-user code you enter there when buying
1.0.0-rc2 · October 25, 2026
Medical, Legal & Tax in English
- Both packs now also work in English, for US and British English. VoxaFlow detects the language of each dictation and uses the matching word lists, writing rules and headings. A German dictation keeps the German rules
- English writing rules for Medical: “fifty milligrams” becomes “50 mg”, “BP one forty over ninety” becomes “BP 140/90 mmHg”, spelled-out abbreviations such as “p r n” become “PRN”. Micrograms become “mcg” in US English and stay written out in British English
- English writing rules for Legal & Tax: “42 U.S.C. § 1983”, “Rule 12(b)(6)”, “[2024] UKSC 12”, amounts in $ or £, percentages and dates by variant (“October 13” or “13 October”)
- Headings in English: “section assessment” or “plan colon”
- 14 English word lists for Medical and 23 for Legal & Tax (US, UK and Latin terms). If an area is on, its English list comes on automatically
- Under Packages you choose US or British English; automatic follows the region of your Mac. The word lists there show each area with its languages
- German writing rules no longer change English text (“19%” used to become “19 %”)
- Guide, practices and getting started with English examples
History
- With Medical or Legal & Tax, the history only saves new dictations if you switch this on under Packages › Professional secrecy. The history page now says so; until now it looked as if the history had stopped working
- “Paste last dictation again” and “Copy last dictation” take the dictation you just spoke. With professional secrecy, an older dictation from the history came instead
1.0.0-rc1 · October 23, 2026
VoxaFlow 1.0
- First release candidate for VoxaFlow 1.0. There are no new features compared to 0.16.2. Until version 1.0 on November 27, only bug fixes follow
- Pro and all packs stay free until 2027-01-31, and updates keep arriving automatically
Legal & Tax
- Latin terms such as “culpa in contrahendo” stay in lower case inside a sentence
0.16.2 · October 22, 2026
Hands-free
- If you hold
Esca little longer, VoxaFlow only pauses and does not discard right after
0.16.1 · October 22, 2026
Hands-free
Escduring a dictation no longer also reaches the app you are dictating into. Some apps, such as browsers and chat apps, used to leave the text field on Esc, and after a pause nothing was inserted
0.16.0 · October 22, 2026
Hands-free
- Pause:
Escpauses a hands-free dictation instead of cancelling it. The dictation key continues,Escagain discards. What you say during the pause is not recorded, and the pause does not count towards the length of the dictation. After ten minutes the pill reminds you once; nothing is discarded on its own - New option under Settings › Keyboard: “One tap starts hands-free”. A short tap is then enough, holding still dictates
0.15.0 · October 22, 2026
Formatting
- Quotation marks by voice: “quote … unquote” or “open quote … close quote” give “…”, in German „Anführungszeichen … Anführungszeichen“, „Anführungszeichen auf … zu“ or „Zitat Anfang … Zitat Ende“ give „…“. Remote Desktop and Citrix get straight quotation marks. A single „in Anführungszeichen“ stays a word
- Subject by voice: if you start a German email with „Betreff“, the rest up to the first full stop becomes the first line, then the email as usual. Copy the first line into the subject field
Guide
- The chapter “Corrections and formatting” explains how to set headings with Medical or Legal & Tax and points to the details
0.14.1 · October 22, 2026
Dictation
- Short sentences like “Was soll ich tun?” are no longer translated into English, for example in Claude or in code editors
- After a sentence end (. ? ! …) VoxaFlow adds a space, so the next sentence after a pause does not stick to the period. Not in terminals
- Long dictations: if the last section cannot be recognized the usual way, it is still recognized instead of going missing
Medical, Legal & Tax
- Headings by voice also work when speech recognition writes them as one word (“Überschriftdiagnose”) or leaves out the period before them (“… Bronchitis Abschnitt Procedere …”). “Im Abschnitt Befund steht …” stays a sentence
0.14.0 · October 21, 2026
Medical, Legal & Tax
- New chapters “Medical package” and “Legal & Tax package” in the guide: set up in four steps, how to dictate, examples for values, dosages, sections of statutes and case numbers, letters with sections, confidentiality, Remote Desktop and Citrix, frequently asked questions
- Medical practice and Legal & Tax practice under Help: five example sentences show in about three minutes what the package writes
- The first time you switch on one of the two packages, a short window shows the first steps. Afterwards the same buttons sit at the top of the package: practice, guide, printable cheat sheet and guide with pictures on voxaflow.ai
- Question marks on the Packages page explain each setting with an example
- Short tips after the first long dictation without a heading and after the first dictation into Remote Desktop or Citrix, each once
Coding
- New chapter “Coding package” in the guide, question marks on the Coding page
Guide
- The guide shows pictures, for example a letter with sections and the Packages page
0.13.0 · October 21, 2026
Medical, Legal & Tax
- Headings by voice: say “Abschnitt Befund”, “Überschrift Procedere” or “Diagnose Doppelpunkt” at the start of a sentence, and VoxaFlow sets the heading. Fixed names count, such as Anamnese, Befund, Diagnose, Beurteilung, Therapie, Procedere and Verlauf, and for Legal & Tax for example Sachverhalt, Rechtliche Würdigung and Begründung. “Der Befund ist unauffällig” stays a sentence
- In documents and emails, for example in Word, Pages, Mail and Outlook, the headings arrive in bold, in Word in the document’s font and size. In practice and office software they appear as “BEFUND:” on a line of their own, in Remote Desktop and Citrix always as plain text. You can change this under Style per app › More …
- Each section is cleaned up on its own; VoxaFlow moves nothing between sections
- The dosage schedule is right at the end of a sentence too: “5 mg 101.” becomes “5 mg 1-0-1.”
Dictation
- In long dictations a correction such as “Nein, sorry, …” or the next list item stays with the sentence before it. The text is there sooner after you release the key, and a correction no longer ends up in the text without what replaces it
- “Hallo, Herr Wagner. Danke …” becomes “Hallo Herr Wagner,” in emails; a “bitte” right after the greeting is written in lower case
- An “M.” that speech recognition writes for “ähm” is dropped at the start of a sentence; “Größe M” or “M wie Martha” stay
Command mode
- “Format this as an email” now sets paragraphs even when the words stay the same. Before, it said “Text unchanged”
License
- The License page no longer shows an update date or an update subscription: a purchase includes all updates of VoxaFlow 1
0.12.5 · October 15, 2026
Packages
- The beta now includes all packages: Coding, Medical and Legal & Tax, without a license key. They stay off until you switch them on under Packages
- The license page and the guide say that all packages are included in the beta and where you switch them on
0.12.4 · October 15, 2026
Connections
- VoxaFlow now asks for the word lists of the packs only if your license includes a pack and you switched it on. Before, the app checked for new lists once a day at launch also without a pack
About and help
- Settings › About › License texts shows the full licenses of the libraries and models used
- The guide lists the requirements under “Getting started”: Mac with Apple silicon, macOS 14, about 3 GB of storage, 16 GB of memory recommended
0.12.3 · October 15, 2026
Medical
- The dosage scheme is right also when speech recognition writes it as one number: “Apixaban 5 mg 101” becomes “Apixaban 5 mg 1-0-1”
- Long technical terms are recognized also when two letters were misheard, for example “Vorhochlimmern” as Vorhofflimmern
- If the cleanup only rephrases a sentence, it is kept instead of falling back to the unprocessed text
Legal & Tax
- File numbers and margin numbers are recognized also when speech recognition joins the digits: “Aktenzeichen 1 BVR 1224, Rand Nummer 15” becomes “Az. 1 BvR 12/24, Rn. 15”
- Ordinary words are no longer replaced by longer technical terms, such as “Rechnung” by “E-Rechnung”
0.12.2 · October 15, 2026
Dictation
- E-mail addresses are written without umlauts and in lower case: “müller@abc.com” becomes “mueller@abc.com”, “ß” becomes “ss”
Connections
- The model download and the update check now send only the app name and version, no longer the macOS version and your Mac’s preferred languages
0.12.1 · October 15, 2026
Remote Desktop and Citrix
- VoxaFlow now types the text into windows of the Windows App and Citrix Workspace, for everyone. This also works when the session blocks the clipboard. Before, Word in the session only received a “v”
- With the Medical or Legal & Tax package you can choose the clipboard instead: faster for long texts, if the session allows it
Dictation
- If music is playing, it goes down as soon as you press the dictation key, not only after about a second. The start sound stays clearly audible
- Spoken e-mail addresses become addresses: “Müller at abc punkt com” becomes “müller@abc.com”
- With content from another device (iPhone, iPad) on the clipboard, a dictation sometimes waited long before the text appeared. This is fixed
Legal & Tax
- “nach 43 StGB” becomes “nach § 43 StGB”, also when “Paragraf” was not recognized
- Amounts are always written in digits: “fünfzig Euro” becomes “50 Euro”, “zwölftausendfünfhundert Euro” becomes “12.500 Euro”
- If a dictated number is missing in the cleaned-up text, you get the text with the simple rules only. What you correct yourself (“um 15 Uhr, nein, um 16 Uhr”) may be left out
0.12.0 · October 14, 2026
Medical and Legal & Tax packages (pilot license first)
- Word lists for Medical (e.g. active substances, laboratory values, cardiology, radiology, nursing, emergency medical services) and for Legal & Tax (e.g. statutes, courts, Latin terms, employment law, tax forms, accounting): “Apixa Ban” becomes Apixaban, “Nicht Zulassungs Beschwerde” becomes Nichtzulassungsbeschwerde
- Dates are written by fixed rule: “am dreizehnten Oktober” becomes “am 13. Oktober”. If the cleaned-up text contains numbers, dates or months you did not dictate, you get the text with the simple rules only
- Professional secrecy: dictations are kept in the history only if you allow it (then for 24 hours, 7 or 30 days at most, on this Mac only), and the history is never synced via iCloud. You can save the sheet “What VoxaFlow sends” as a PDF for your record of processing activities
- Style per app has the categories “Practice software” (medical letter, findings and notes, short note) and “Law and tax office software” (brief and client letter, file note, short note). You choose which apps belong to them on the Packages page
- Remote Desktop and Citrix: VoxaFlow inserts into the window of the Windows App or Citrix Workspace, per client via the clipboard or by typing if the session blocks the clipboard
Packages
- “Check now” shows new word lists right away, not only after a few minutes
0.11.1 · October 14, 2026
Long dictations
- After release the text arrives faster: for a one-minute dictation on an M4 in about half a second, before sometimes over a second
- Enumerations (“I have two questions. Firstly … secondly …”) reliably stay a list in long dictations too
0.11.0 · October 14, 2026
Coding
- You switch packages on as a whole on the “Packages” page. The Packages page only appears when your license includes a package, and shows only your packages
- With the Coding package switched on, the new “Coding” page appears in the VoxaFlow window. If you already used Coding word lists in 0.10, the package is already on
- Projects: when you dictate in Cursor, VS Code, Xcode, Zed, JetBrains or a terminal inside a git repository, VoxaFlow knows the names from the project. “auth service” becomes
AuthService, “fetch user profile” becomesfetchUserProfile. It is read on your Mac only - Each project can have its own dictionary entries and snippets, which come before the global ones there
- Terminal: when a dictation starts with a command, it comes without AI and without a period: “git commit minus m fix login bug” becomes
git commit -m "fix login bug". If an agent such as Claude Code runs, a sentence stays a sentence - Prompts in English: speak German, write English, switched on separately for editors and terminals. Names from the project, file names, paths and commands stay unchanged
- Spoken casing in every app: “camel case user id” →
userId, “snake case user id” →user_id, “constant max retries” →MAX_RETRIES, plus pascal and kebab case
0.10.0 · October 13, 2026
Packages
- New “Packages” page in the VoxaFlow window. There you switch on word lists one by one, for Coding for example React, Next.js, TypeScript, Swift, Python, Kubernetes, Supabase or AI tools
- While you dictate, VoxaFlow recognizes terms from the lists you switched on, also when they sound alike or were recognized slightly differently: “use effect” becomes
useEffect, “super base” becomes Supabase, “kube control” becomeskubectl. Your own dictionary comes first, and ordinary words are not replaced - During the beta the Coding lists are free. Once a day VoxaFlow looks for new versions; you can switch that off on the page
- The Medical and Legal & Tax packages come with writing rules: “fünf Milligramm” becomes “5 mg”, “eins null eins” becomes “1-0-1”, “Paragraf achthundertdreiundzwanzig Absatz eins BGB” becomes “§ 823 Abs. 1 BGB”. At first they are only available with a pilot license
Dictation
- When speech recognition hears an “ähm” as “im” before “ich”, “wir” and similar words, it is now reliably removed
0.9.0 · October 13, 2026
Output language and own instruction per app
- Settings › AI › Style per app now has “More …” for every kind of app. There you choose an output language, for example speak German and write English in chat
- You can also add an own instruction of up to 300 characters, such as “End every email with: Best regards, Anna” or “no exclamation marks”
Own keys for hands-free and command mode
- Under Settings › Keyboard › Keys per mode you set an own hands-free key and an own command key. Tapping twice and ⌃ Control with the dictation key keep working
Long dictations
- A dictation can last up to 15 minutes. From 14:30 the pill counts down the seconds; at 15:00 the recording ends and the text is inserted
Free and Pro
- Features are now assigned to Free and Pro. During the beta everything stays free until January 31, 2027
- From version 1.0 on, VoxaFlow starts with 14 days of Pro to try. After that, without a license, the Free limits apply: 2,500 words a week, 2 minutes per dictation, 25 dictionary terms and 7 days of history. Saved terms, snippets and older dictations are kept
0.8.2 · October 13, 2026
Command mode
- In browsers VoxaFlow now reliably finds selected text, also in Chrome, Edge, Arc and other Chromium browsers and in fields that web pages build themselves
- If VoxaFlow finds no selection, it inserts nothing and says “No selected text found”. Before, the spoken instruction was inserted as a dictation and replaced the selected text
Dictionary suggestions
- Corrections in Outlook, webmail and other rich-text editors are now recognized
- You have up to five minutes after inserting to make corrections, even if you switch to another app in between
- Sent or emptied messages no longer count as “changed by you” in the statistics
Updates
- While an update is waiting, the menu bar icon shows a dot, and the update window comes up at most once a day, never during a dictation. Before, an update mostly showed only in the menu
Audio
- Other audio such as music or videos is turned down during the recording instead of muted, and set back to the previous volume afterwards. The setting is now on by default (Settings › Audio)
- If VoxaFlow crashes during a recording, the volume is back to normal at the next launch
Menu
- “Open VoxaFlow …” (
⌘,) replaces “Settings …” and opens the window at the top of the overview, even if it is already open
iCloud sync
- The “History” switch now applies to all your Macs. Before, it had to be turned on on each Mac, otherwise no history was transferred
- The Macs no longer rewrite their data every minute when nothing has changed
0.8.1 · October 8, 2026
License
- “Buy” on the License page reported a license server error during the beta. Sales start on January 1, 2027; until then the page says so, and the beta stays free with all features
0.8.0 · October 8, 2026
License
- New “License” page in the VoxaFlow window: it shows until when the beta runs and, after buying, your license with packages, feature updates, update subscription and devices
- Enter and activate the license key from the purchase email; case, spaces and mixed-up characters such as O and 0 do not matter
- “Deactivate this Mac” frees the seat for another Mac
- Buy packages, the update subscription or a new license right on the page; the checkout opens in your browser, and beta testers get the launch price
- The end of the beta now comes from the license server. VoxaFlow checks every 7 days, with a license every 30 days. It only sends a checksum of the hardware ID, the app version and the key or license token, never text or audio. Without internet everything keeps working
- The menu only shows a notice when the license needs your attention, for example when it could not be checked for a long time
0.7.6 · October 8, 2026
Menu
- During the beta “Send Feedback …” is right in the menu, so problems and wishes are quick to send. After the beta it moves into the Help submenu
- The duplicate item “Open VoxaFlow …” is gone: “Settings …” opens the same window, with the overview, history and dictionary in its sidebar
0.7.5 · October 8, 2026
- VoxaFlow no longer adds a greeting or closing you did not dictate. Until now “Guten Tag” and “Viele Grüße” often ended up around the text, even when you dictated just one paragraph into an existing email. Dictated greetings and closings stay, on their own lines as before
- In command mode, “format this as an email” no longer adds a greeting or closing either; if you want one, say for example “add a greeting and a closing”
Menu
- The menu bar menu is tidied up: one status line at the top, then the last dictation, with “History” and “Help” as submenus
- Your own shortcuts appear next to the items
- “Settings …” and
⌘,always open the General page, even when the window is already open - The notice about the end of the beta only appears in the menu in the last two weeks
Feedback
- New under Help › Send Feedback …: send problems, ideas and wishes straight from the app, optionally with an email address for questions. For problems VoxaFlow attaches a diagnostic report without dictated text, which you can view first. Nothing is sent until you click Send; without internet it goes by email. Replaces “Report a Problem …”
0.7.4 · October 8, 2026
Pasting
- On web pages, dictation is pasted directly again, for example in a post or comment in a Skool community. Some pages report their input field as a button, so the text wrongly ended up only on the clipboard with the “No text field” notice
Updates
- If you skipped updates, the update window shows every change since your version, grouped by version. Until now it showed only the newest version
0.7.3 · October 7, 2026
Email format
- Dictated emails arrive structured: greeting on its own line with a blank line after it, paragraphs in the body, a blank line before the closing, closing and name on their own lines. German closings like “Mit freundlichen Grüßen” no longer get a comma (DIN 5008)
- Works in Mail, Outlook, Spark and other mail apps and now also in webmail in the browser (Gmail, Outlook on the web, GMX, WEB.DE and more), recognized by the window title. The email style applies there too
- New command: select text and say “format this as an email”
- Can be switched off under Settings › AI › Style per app › Email format
Command mode
- Select text, start command mode and say “rewrite this”, “shorter”, “more formal”, “make it a bullet list” or “fix the mistakes”: VoxaFlow now really changes the selected text. Before, many instructions other than translating returned the text unchanged
- Dictionary terms no longer appear below the text
- If the text still comes back the same, the selection stays as it is and the pill shows “Text unchanged. Try a different instruction.”
0.7.2 · October 6, 2026
Pasting
- In Microsoft Word and other apps with their own editor, dictation is pasted directly again. Since 0.7.0 it wrongly ended up only on the clipboard with the “No text field” notice
0.7.1 · October 6, 2026
New installer window
- The installer disk image window now shows the VoxaFlow design: dark background with the waves from voxaflow.ai, the logo, a clear instruction in English and German and an arrow from the VoxaFlow icon to the Applications folder
- The mounted disk shows the VoxaFlow icon
New welcome window
- Setup on first launch runs in four steps: welcome, permissions, dictation key, models. With moving waves in the background that swell when you press the dictation key
- The dictation key can be changed right in the welcome window, e.g. for external keyboards without fn
- New: “Open VoxaFlow at login”, switched on by default, so the dictation key works right after every restart. Change it under Settings › General
- With “Reduce motion” in the macOS accessibility settings the waves stand still
0.7.0 · October 6, 2026
New main window
- Instead of the tab bar, a window like System Settings: sidebar with search, colored icons and a header card per section, back and forward. New sections Overview, Inserting, iCloud & Data and Permissions; “Hotkey” is now “Keyboard”
- While the window is open, VoxaFlow shows in the Dock and in ⌘-Tab
- Open at login (Settings › General)
Overview
- Words dictated, time saved compared with typing (typing speed adjustable, 40 words per minute by default), dictations, speaking speed, snippets used, each compared with the previous period
- Corrections split into AI cleanup, dictionary and “by you afterwards”
- Words per day, activity by weekday and hour, top apps. Numbers only, never the dictated text, with its own switch under iCloud & Data
History and copying
- History with search, filter by period and app, groups by day and app icon. Copy by double-click,
⌘Cor the icon in the row, “Copy original”, “Clean up again” in another style, export as text, CSV or JSON - Retention of 7 days, 30 days, 1 year or forever instead of a fixed 500 entries
- Menu bar: “Paste last dictation again” and recent dictations as a submenu, also after a restart
- If no text field is active (e.g. on the desktop), the text stays on the clipboard and the pill shows “No text field”. Before, it was lost
Shortcuts and quick history
- Your own key combinations for quick history, “Paste last dictation again”, “Process last recording again” and “Add selection to dictionary”, without extra permissions
- Quick history as a floating window like Spotlight: search,
↩inserts into the app you are writing in,⌘Ccopies,⌘↩inserts the original - The last recording stays in memory and can be processed again, e.g. after an error
Dictionary and snippets
- Table with search and sorting, edit by double-click, duplicate terms are merged
- Import and export as JSON or CSV (Excel, Numbers) with a preview “new, updated, unchanged”, merge or replace
- Learning from corrections: when you change a word right after inserting, VoxaFlow suggests it for the dictionary. Only the text field VoxaFlow wrote into is read, never a password field; nothing is added without asking
- Backup and restore of settings, styles, dictionary, snippets and optionally the history in one file
iCloud sync
- Dictionary, snippets, settings and statistics on all Macs, and the history if you want. Stored in the “VoxaFlow” folder in iCloud Drive
- End-to-end encrypted; the key lives in the keychain, further Macs get it through a recovery code. Neither Apple nor VoxaFlow can read the data
- Changes apply per entry, deletions reach all Macs, the same term from two Macs becomes one. If iCloud Drive is off, VoxaFlow says so
Audio
- Optionally mute other audio (music, videos) while recording; afterwards everything is as before
- Fallback microphones in a fixed order when the chosen microphone is not connected
0.6.6 · October 5, 2026
Beta until January 31, 2027
- The menu, the welcome window and Settings › About show “Beta: all features, free until January 31, 2027.”
- After that VoxaFlow says once that the beta has ended and points to voxaflow.ai; the menu shows “Beta ended: get the current version …”. Dictation keeps working
Model download
- An interrupted download of the language models resumes where it stopped on the next attempt, even after a restart; before, it started over
- Every file is checked after downloading, a damaged one is downloaded again. VoxaFlow retries short connection drops by itself
- A downloaded model starts without an internet connection
- Models you already downloaded are taken over, without a new download and without extra disk space
- Clear messages when there is no internet, a server error or a full disk, plus “Try again” in the welcome window, the settings and the menu. When the network is back, VoxaFlow continues by itself
Report a Problem
- New in the menu: “Report a Problem …”. VoxaFlow creates a text file with version, Mac model, settings, log entries of the last 24 hours and the latest crash report, and opens an email to hello@voxaflow.ai with the file attached. Alternatively just save the file
- No dictated text, no audio, no history; dictionary and snippets only as a count. Nothing is sent until you click Send yourself
Faster on Macs with M1
- On the M1, long dictations are cleaned up about twice as fast (1.4 to 2.4 instead of 4 to 8 seconds), short ones about 30 % faster. The large model Qwen3 4B is now as fast there as the small one, with the same accuracy
0.6.5 · October 3, 2026
Updates without interrupting windows
- When the daily check finds an update in the background, VoxaFlow no longer opens a window over your work but shows “Install update …” in the menu bar. Only when the check runs right after launch or you check yourself does the update window appear immediately
0.6.4 · October 3, 2026
Automatic updates
- VoxaFlow checks once a day for a new version and installs it on request. New in the menu bar: “Check for Updates …”; under Settings › About: “Check for updates automatically” and “Check now”
- Updates come from
updates.voxaflow.ai(servers in Germany), are signed and notarized by Apple - From this version on updates arrive automatically; from 0.6.3 and earlier install the DMG once by hand
Models from our own server
- Speech recognition and language models come from
models.voxaflow.aiinstead of Hugging Face, in fixed versions - Models you already downloaded stay valid, nothing is downloaded again
0.6.3 · October 3, 2026
Long dictations: processed while you speak
- From 8 seconds of recording on, VoxaFlow recognizes and cleans up finished sentences while you keep talking. After you release the key, the text usually appears after about a second, even after a minute and a half of dictation (before 4 to 5 seconds)
- If you correct yourself at the start of a new sentence (“no”, “sorry”) or continue a list (“secondly”), that sentence is cleaned up together with the previous one
- Fixed: periods at pauses in the middle of a sentence and a dropped subordinate clause
Filler words
- After the AI cleanup VoxaFlow removes remaining filler words (uh, um, erm, hm and others; “um” only in English text). An inserted filler word takes its commas with it
- Fixed: without a language model, “um” was removed in German too (“um 16 Uhr” → “16 Uhr”)
- “nee” counts as a correction word
0.6.2 · October 3, 2026
Fixed microphone
- Fixed: with a fixed USB microphone (e.g. Jabra Evolve Link MS) every recording stopped after about half a second with “No microphone found”
Start sound only when really recording
- Bluetooth headsets deliver only silence at first while switching. The start sound now plays only once the microphone actually records, so the first word is not lost
More accurate models
- Large model: a more accurate version of Qwen3 4B, with the same size and speed
- Small model: Qwen3.5 2B replaces Qwen3 1.7B, more accurate, faster on long dictations and with less memory
- On the first launch after the update VoxaFlow downloads the new model (2.3 or 1.7 GB)
Faster short dictations
- Short dictations are cleaned up about 0.1 seconds faster
Small model for Macs with little memory
- New: a second, smaller local model, about twice as fast but somewhat less accurate
- Without your own choice the model follows the memory: under 12 GB the small one, otherwise the large one. The large one stays selectable; on Macs with little memory the settings and the welcome window point out that recognition can get slower, with a button to switch
- In Low Power Mode the AI cleanup takes about twice as long; the guide now says so
Command mode: key order no longer matters
- Fixed: pressing
fnfirst and then⌃ Controlgave a normal dictation, and the spoken instruction replaced the selected text - If
⌃ Controlfollowsfnwithin a second, the recording switches to command mode (the pill turns purple)
0.6.1 · October 1, 2026
Permissions stay stable
- Permissions are kept across updates and reinstalls
- Input Monitoring is no longer needed; VoxaFlow needs only the microphone and Accessibility
- New button “Repair permissions”: removes stale entries and requests the permissions again; working permissions stay untouched
- If VoxaFlow was installed with a different signature, the app explains it instead of just saying “not active”
Notarized by Apple
- VoxaFlow is signed with Developer ID and notarized by Apple and opens without a Gatekeeper warning
0.6.0 · October 1, 2026
Settings reorganized
The General tab had grown too long. The settings now have eight tabs:
| Tab | Contents |
|---|---|
| General | App language, clipboard, permissions |
| Hotkey (new) | Dictation key, usage hint, key check with conflicts and key test |
| Audio (new) | Devices (microphone, sound output), sounds (on/off, start sound) |
| AI (new) | Speech recognition (dictation language), cleanup (model with recommendation), style per app, local models |
| Dictionary, Snippets, History, About | unchanged |
- The former Style tab is now the “Style per app” section in the AI tab
- The menu bar opens the matching tab: “fn key is assigned twice” → Hotkey, “Set up permissions” → General
Guide
- New chapter “Settings”: overview of the eight sections and the menu bar items
- “Troubleshooting”: new sections “The text appears in the wrong language” and “The start sound is missing or annoying”
- “?” help also for app language, audio devices and start sound
0.5.0 · October 1, 2026
Audio devices
- New in the settings: choose the microphone and sound output, “System default” or a fixed device
- If the fixed device is not connected, the system default applies automatically, with a notice in the settings; when it comes back, it is used again
- Fixed: after switching from AirPods to a USB microphone, VoxaFlow recorded nothing
- If AirPods switch to headset mode when recording starts, the recording continues without losing what was said
- If the microphone delivers no data, VoxaFlow restarts the recording once, then shows “Microphone not responding”
- Tip for Bluetooth microphones: choose the Mac’s microphone so the AirPods don’t switch to headset mode
Start sound
- “Glass” removed, new last option “No start sound” (the start is then shown only by the pill)
Language is kept
- Fixed: a German dictation into a code app such as Claude was translated into English. The cleanup now always keeps the language of the dictation (except in command mode, where translating is allowed)
0.4.0 · October 1, 2026
Bilingual (English and German)
- The whole interface is available in English and German: menu bar, settings, welcome window, pill, messages, guide and guided practice
- The language follows macOS; English applies for all other system languages
- New in the settings: “App language” (system, Deutsch, English), effective after a restart with a “Restart now” button
- “Language” for speech recognition is now called “Dictation language”
Sounds
- Fixed: the start sound was missing on the first press and after a longer pause
- New default start sound “VoxaFlow sound”, a short, soft tone matching the end sound. Before, the start sound sounded like the macOS alert
- Start sound selectable (VoxaFlow sound, Bottle, Purr, Morse, Glass, Tink) with a preview button
Silence no longer produces text
- Fixed: pressing and releasing the key without speaking inserted made-up filler words. Without speech nothing is inserted now, the pill briefly shows “Nothing recognized”
- Silence at the start and end of a recording is trimmed
- Formal emails no longer replace “thanks” with “Thank you”
0.3.0 · October 1, 2026
Faster
- The AI cleanup of typical dictations takes 0.2 to 0.6 instead of 1.0 to 1.4 seconds
- VoxaFlow prepares the cleanup for the active app as soon as you press the dictation key
Qwen3 4B recommended
- Qwen3 4B was faster and more accurate than Apple Intelligence in every test and is now marked as recommended in the settings. With another choice there is a notice and a “Switch to Qwen3 4B” button, also in the welcome window
- Guide: new section “Which language model?” with a comparison
Changed
- The style per app changes only punctuation, capitalization and line breaks, not your wording
Known limitations
- Long dictations still take more than a second
0.2.0 · October 1, 2026
New
- Guide in its own window: 10 chapters with before/after examples, tips and practice fields, in English and German
- Guided practice in 5 steps (dictation, corrections, lists, hands-free, command mode); VoxaFlow ticks off each step automatically
- “?” buttons and tooltips in the settings with an explanation, an example and a jump to the matching chapter
- Entry points in the menu bar (“Guide …”, “Guided practice …”), the welcome window (“Start practice”) and Settings › About
Fixed
- Command mode also works in VoxaFlow’s own windows, such as the practice fields
0.1.1 · October 1, 2026
Fixed
- The fn check wrongly reported “Press 🌐 twice” as a conflict although Apple’s dictation was off
0.1.0 · October 1, 2026
First working version of the Mac app.
Dictation
- Push-to-talk with fn (alternatively right ⌥ or right ⌘), hands-free mode by tapping twice, Esc cancels
- Command mode: ⌃ + fn with selected text rewrites the selection by voice instruction
- Text is inserted at the cursor, the previous clipboard content is restored
- Pill with level meter, menu bar app with its own icon
Local AI
- Speech recognition with Parakeet TDT v3 on the Neural Engine
- Text cleanup with Qwen3-4B-Instruct: filler words, self-corrections, punctuation, tone depending on the app
- Alternatively Apple Intelligence (macOS 26 and later) or simple rules only
- Lists (“firstly … secondly …”) become numbered lists
Personalization
- Dictionary with correction of misrecognitions, snippets, style per app category, local history
Setup
- Welcome window with permissions and model status; detects granted permissions without a restart
- Check whether fn/🌐 is assigned elsewhere, with a warning in the settings, the menu bar and once in the pill
- Live key test for the dictation key
Known limitations
- The AI cleanup takes about 1.1 to 1.4 seconds
- Only the 25 languages of Parakeet
- Not notarized yet, no auto-update
VoxaFlow for Windows
1.0.0-rc7 · November 16, 2026
Fixed from feedback
- Voices in the background: when someone in the same room spoke, say a few metres away, their words sometimes ended up in the middle of your dictation. VoxaFlow now leaves out speech that is much quieter than your own voice before recognition. You can turn this off in Audio with Ignore quiet voices in the background. When two people talk at the same time, their voices cannot be separated.
- Guided practice: dictations sometimes did not appear in the practice field, mostly right after setup, even though the practice moved on. Now the text lands in the field, and the field is ready for dictation straight away, also on the last page of setup. Thanks for the report!
- VoxaFlow spells its own name correctly, even when the recognizer hears something else, such as “Vuxaflo”. You don’t need a dictionary entry for it.
1.0.0-rc6 · November 10, 2026
New: notes
- The main window has a Notes section: write notes or speak straight into them with Dictate, search them, copy them and pass them on with Share
- Notes stay on this PC and are part of the backup. A backup from the Mac brings its notes along
1.0.0-rc5 · November 4, 2026
Fixed from a user report
- English words in the middle of a German dictation: speaking German quickly with many English technical terms, for example when dictating into Claude Code, sometimes produced half-English sentences. The speech recognition heard German words as similar-sounding English ones, such as “by” for “bei”, “so that” for “sodass” or “here any” for “hier keine”, and the AI cleanup left them in. With a “not” among them, VoxaFlow even inserted the dictation without any cleanup. The AI cleanup now brings such words back to German; technical terms, names and code stay. Thanks for the report!
AI cleanup
- If you dictate an English sentence within German, such as a commit message, a quote or a code comment, it now stays in English word for word. VoxaFlow Lektor sometimes translated such sentences into German along with the rest (“fix the login bug” became “Fixe den Login-Bug”). If the model translates anyway, VoxaFlow inserts the text with the basic cleanup only
1.0.0-rc4 · November 3, 2026
VoxaFlow Lektor
- New models for the AI cleanup: VoxaFlow Lektor replaces Qwen3 4B, VoxaFlow Lektor Kompakt replaces Qwen3.5 2B. Both are built on Qwen3, trained for dictation and, as before, run only on your PC
- Lektor Kompakt makes far fewer mistakes than Qwen3.5 2B and takes about half as long as VoxaFlow Lektor
- Your choice stays: if you had Qwen3 4B, you get VoxaFlow Lektor; if you had Qwen3.5 2B, Lektor Kompakt. “Automatic” still picks what suits your PC
- After the update, VoxaFlow downloads the new model in the background (2.5 or 1.1 GB). Until it is ready, the previous model keeps cleaning up. Then VoxaFlow deletes the old models and frees the space
AI cleanup
- If you correct yourself with “no, not …, but …”, the corrected version comes through: “cheese, no, not cheese, but butter” becomes “butter”. Before, VoxaFlow often inserted such dictations without the cleanup
- Numbers you say digit by digit, such as “zero one seven six, two three four …”, come through as figures. Before, VoxaFlow often inserted such dictations without the cleanup
- If you announce a list, such as “1. … and secondly …”, it comes through as a list. Before, VoxaFlow inserted such dictations without the cleanup
- Choosing another AI model on the AI page takes effect right away, also shortly after the start. Before, the switch could take several minutes or only work after a restart
1.0.0-rc3 · October 28, 2026
Build 2 of October 28, 2026 brings the fixes from the feedback of October 27, 2026; build 3 of the same day shows what’s new when updating, recognizes e-mail addresses more reliably and brings the early-user code for the launch price.
Updates
- Before an update VoxaFlow shows what it brings: a window lists what’s new since your version, with “Restart now” or “Later”. It comes up at most once a day and never during a dictation
- If the update was installed at the start, what’s new shows once at the first start of the new version. You can find it any time under Help › What’s new in the VoxaFlow menu and on the About page
AI cleanup
- A “please” or a negation such as “not” or “never” is no longer lost in the cleanup. Before, when a sentence had two “bitte” (German “please”), the second one was often missing. If the AI still leaves one out, VoxaFlow inserts the text without its cleanup
- With an NVIDIA graphics card with 6 GB or more, “Automatic” now runs the AI on the graphics card, mostly in under a second. Other graphics cards stay on the processor until they have been measured
Inserting
- When the cursor is in no text field, for example on a button, in a file list or on a web page without an input field, VoxaFlow no longer pastes into nowhere. The text stays on the clipboard, Ctrl+V pastes it. Before, the old clipboard came back afterwards
- Spoken e-mail addresses are recognized more reliably, also when speech recognition breaks them up: “Holger at Login minus Online Punkt com”, “holger atlogin-online.com” or “Kanzlei Punkte” (for “Punkt de”) give a correct address. An address at the start of a sentence stays in lower case
- The last word of a dictation is lost less often, for example the “de” at the end of an address
License
- The License page, menu, setup and guide now speak of the free period until January 31, 2027. Everyone who uses VoxaFlow before the sales start gets the launch price
- The License page shows your early-user code for the launch price. Everyone who uses VoxaFlow before February 1, 2027 gets the launch price on every purchase until February 28, 2027: on this PC automatically, on another computer with the early-user code you enter there when buying
Dictation key
- A shortcut with the dictation key, such as Ctrl+P with the right Ctrl key, no longer starts a dictation you can hear: no start sound, no pill, no “Cancelled”. Sound and pill only come once the key is held for about a quarter of a second
Guide
- The Medical and Legal & Tax chapters now give the prices with the new sales start: from February 1, 2027
1.0.0-rc2 · October 26, 2026
License
- Sales start on February 1, 2027 instead of January 1. Until then Pro and all packages stay free, until January 31, 2027, also for new installations
1.0.0-rc1 · October 26, 2026
VoxaFlow for Windows 1.0
- First release candidate for Windows. The Windows version now does everything the Mac version does except iCloud sync, and carries the same version number. Until the final version only bug fixes follow
- Pro and all packages stay free until January 31, 2027; updates keep coming automatically
Settings
- Under AI the performance level of this PC (Recommended, Supported, Limited or Not supported) shows with its reason and a link to the computer check. On Windows 10 and Windows on ARM VoxaFlow says once why it is not supported there
0.1.19 · October 26, 2026
Data
- Backup: settings, styles, dictionary, snippets and optionally the history in one file, e.g. for a new PC. When restoring you choose the parts; dictionary, snippets and history are merged. A backup from the Mac brings its dictionary, snippets and history
History
- Clean up again: have a dictation cleaned up once more, in the same or another style
Packages
- Word lists now come signed from packs.voxaflow.ai, as on the Mac, and are checked for new versions once a day. At the first start after the update VoxaFlow loads the lists of the packages that are on. Under Packages › Word lists you can check now or remove the lists
- “What VoxaFlow sends” as a PDF for a practice’s or office’s record of processing activities (Packages › Professional secrecy)
License
- The license page shows the scope as on the Mac and, from the sales start on January 1, 2027, buying and adding packages. During the beta everything stays free until January 31, 2027
About
- License texts of the models and software used
- Checking for updates automatically can be switched off
0.1.18 · October 26, 2026
Coding
- Project dictionary: when you dictate in a Git repository, VoxaFlow reads its file and folder names and the names of types and functions once. “auth service” becomes AuthService, “fetch user profile” becomes fetchUserProfile, “auth service dot swift” becomes AuthService.swift. Files ignored by Git are skipped, and nothing of it leaves your computer
- VoxaFlow finds the project in the window title: by the folder name in VS Code, Cursor and Windsurf, by the path in Command Prompt and Git Bash. Add other folders under Coding › Projects
- Each project can have its own dictionary entries and snippets that come before the global ones there
0.1.17 · October 25, 2026
Updates
- “Restart now” under About only worked on the second click and then showed a plain window. One click is enough now; the update installs without a window and VoxaFlow restarts by itself
0.1.16 · October 25, 2026
History
- With Medical or Legal & Tax, the existing history was deleted and the history page stayed empty. Older dictations are now kept. New ones are still only saved if you switch this on under Packages › Professional secrecy; the history page now says so
- “Paste last dictation again” and “Copy last dictation” take the dictation you just spoke, even when professional secrecy keeps it out of the history
0.1.15 · October 25, 2026
Setup
- The progress bar while the models load for the first time flickered and kept starting over. It now grows steadily
0.1.14 · October 25, 2026
Bar at the screen edge
- The bar at the bottom of the screen sometimes stayed visible after inserting. It now disappears reliably
0.1.13 · October 23, 2026
Medical, Legal & Tax in English
- Both packs now also work in English, for US and British English. VoxaFlow detects the language of each dictation and uses the matching word lists, writing rules and headings. A German dictation keeps the German rules
- English writing rules for Medical: “fifty milligrams” becomes “50 mg”, “BP one forty over ninety” becomes “BP 140/90 mmHg”, spelled-out abbreviations such as “p r n” become “PRN”. Micrograms become “mcg” in US English and stay written out in British English
- English writing rules for Legal & Tax: “42 U.S.C. § 1983”, “Rule 12(b)(6)”, “[2024] UKSC 12”, amounts in $ or £, percentages and dates by variant (“October 13” or “13 October”)
- Headings in English: “section assessment” or “plan colon”
- 14 English word lists for Medical and 23 for Legal & Tax (US, UK and Latin terms). If an area is on, its English list comes on automatically
- Under Packages you choose US or British English; automatic follows the Windows region. The word lists there show each area with its languages, one switch for all of them
- German writing rules no longer change English text (“19%” used to become “19 %”)
- Practices, the package window and tips with English examples
Legal & Tax
- Latin terms such as “culpa in contrahendo” stay in lower case inside a sentence
0.1.12 · October 22, 2026
Hands-free
- Esc for pausing or cancelling no longer also reaches the program you dictate into. Some programs, such as browsers or Teams, left the text field on it, and the text was not inserted after the pause
- Holding Esc a little longer only pauses and no longer discards right after
0.1.11 · October 22, 2026
Hands-free
- Esc pauses a hands-free dictation instead of cancelling it. The dictation key continues, Esc again discards. What you say during the pause is not recorded, and the pause does not count towards the length of the dictation
- New option under Settings › Keyboard: “One tap starts hands-free”. A short tap is then enough; holding still dictates
Formatting
- Spoken quotation marks: “quote … unquote” or “open quote … close quote” give “…”, in German „Anführungszeichen … Anführungszeichen“, „Anführungszeichen auf … zu“ or „Zitat Anfang … Zitat Ende“ give „…“. Remote Desktop and Citrix get straight quotation marks
- Subject by voice: if you start a German email with „Betreff“, the rest up to the first full stop becomes the first line, with the email below as usual
0.1.10 · October 22, 2026
Updates
- When an update is ready, the VoxaFlow icon in the notification area carries a dot, and the pill says so once. Restart from the icon’s menu, or the update is applied at the next start
- VoxaFlow applies updates without a window of its own: the pill announces the restart, and shortly after the new version starts by itself. This applies from the next update after 0.1.10
- The installer shows the VoxaFlow logo
0.1.9 · October 22, 2026
Dictation
- After Windows started, the dictation key on some PCs only worked once VoxaFlow had been clicked. VoxaFlow now checks on its own whether it still detects the key, reconnects if needed and starts the dictation anyway
- If you hold the dictation key in a program running as administrator, the pill shows a hint: Windows does not let VoxaFlow detect keys there
- Short sentences like “Was soll ich tun?” are no longer translated into English, for example in Claude or in code editors
- After a sentence end (. ? ! …) VoxaFlow adds a space, so the next sentence after a pause does not stick to the period. Not in terminals
- Long dictations: if the last section cannot be recognized the usual way, it is recognized plainly instead of the whole dictation being lost
History
- In a narrow window the search and filters covered the Export and Delete buttons. The buttons are now above the text and wrap into a second line when needed
0.1.8 · October 21, 2026
Speed
- On battery, VoxaFlow sets the Windows power mode to “Best performance” during a dictation and back afterwards. The AI then works about twice as fast as in “Best power efficiency”; in the test a long dictation was done after 2.9 instead of 5.0 seconds. Not in energy saver; switch it off under Settings › AI
0.1.7 · October 21, 2026
Medical, Legal & Tax
- Headings by voice also work when speech recognition writes them as one word (“Überschriftdiagnose”) or leaves out the period before them (“… Bronchitis Abschnitt Procedere …”). “Im Abschnitt Befund steht …” stays a sentence
- When you switch Medical or Legal & Tax on for the first time, a short window shows examples and links to the guide with pictures and the cheat sheet. The same links then appear under “Getting started” at the top of the package, and the guide for Coding
- Two tips appear once each in the pill: on the first dictation into a remote session, how the text gets there, and after the first long dictation without a heading, how to start a section
Help
- The guide is now in the app: every chapter with examples and practice fields, written for Windows and with your dictation key. Open it under Help in the notification area, under About and under “How to dictate”
- Guided practices: five short steps for dictating, correcting, hands-free and command mode, and with Medical or Legal & Tax a practice of its own. VoxaFlow checks off each step as soon as the dictation is right. You can also start the practice at the end of setup and in the window that appears when you switch a package on for the first time
- Many settings have a “?” next to them: one click explains the setting and leads to the matching chapter
- The guide on voxaflow.ai still opens under Help in the notification area and under About; with a package on, also the Medical or Legal & Tax guide
0.1.6 · October 21, 2026
Medical, Legal & Tax
- Headings by voice: say “Abschnitt Befund”, “Überschrift Procedere” or “Diagnose Doppelpunkt” at the start of a sentence, and VoxaFlow sets the heading. Fixed names count, such as Anamnese, Befund, Diagnose, Beurteilung, Therapie, Procedere and Verlauf, and for Legal & Tax for example Sachverhalt, Rechtliche Würdigung and Begründung. “Der Befund ist unauffällig” stays a sentence
- In documents and emails, for example in Word and Outlook, the headings arrive in bold, in Word in the document’s font and size. In practice and office software they appear as “BEFUND:” on a line of their own, in Remote Desktop and Citrix always as plain text. You can change this under AI › Style per app › More …
- Each section is cleaned up on its own; VoxaFlow moves nothing between sections
0.1.5 · October 18, 2026
Dictation
- Learning from corrections: if you correct a word right after it was inserted, VoxaFlow suggests it for the dictionary. Suggestions appear at the top of the dictionary; you can turn this off under Insert. VoxaFlow only reads the field it has just written into, never a password field
Controls
- Custom shortcuts for quick history, insert last dictation again, reprocess last recording and add selection to dictionary (Keyboard › Shortcuts)
- The app is now available in English: General › App language, automatically following the Windows display language
- The pill reliably appears above the window you are dictating into again. After an update it was sometimes lost behind other windows
About and license
- About says beta instead of prototype
- Instead of an “updates until” date, the license page shows that all updates of VoxaFlow 1 are included
0.1.2 · October 18, 2026
Speed
- Dictations no longer wait a long time after you change packages, the dictionary or other settings. Until now the text took 10 to 22 seconds longer afterwards. A 36-second dictation is now ready 3.3 instead of 15.6 seconds after you release the key
0.1.1 · October 18, 2026
Startup
- VoxaFlow starts properly after installation. 0.1.0 was missing prompts and word lists, so the app kept running without a window or a notification-area icon. If you installed 0.1.0, install 0.1.1 once with the installer, because 0.1.0 never got as far as checking for updates
- If startup fails, VoxaFlow shows a message
0.1.0 · October 17, 2026
First beta for Windows 11
- Speech recognition (Parakeet) and editing (Qwen3) run entirely on your PC, without the cloud
- The dictation key is the right Ctrl key: hold it and speak, tap it twice for hands-free mode, Esc cancels. Shift and the right Ctrl key start command mode for selected text
- As on the Mac: dictionary, snippets, style per app, history and the Coding, Medical and Legal & Tax packages (all included in the beta), also for Remote Desktop and Citrix
- VoxaFlow sits in the notification area of the taskbar. A click opens the main window with all settings, a right-click opens the menu with history and help
- The setup assistant downloads the models and picks the language model that suits your PC: Qwen3 4B, Qwen3.5 2B or rules only
- VoxaFlow downloads updates in the background and installs them at the next start
- Not included yet: iCloud sync, the project dictionary for Coding, the in-app guide and the guided practice