Fact ยท Dictation software
236sources
The features of dictation and voice typing apps dataset draws on 236 distinct sources.
Calculated from 602 data points on October 19, 2026
- How it is calculated
- Count of the distinct sources behind the dataset's 602 rows: 236.
A statistic of the dataset Dictation App Features Compared: Offline, Privacy, AI, calculated by code from its published rows.
Calculated from
- SpeechAgent's page says the extension has no server-side code.
- SpeechAgent's page says the speech recognition occurs on a server, not on your device.
- SpeechAgent's page says it works in text areas on the active web page, and that elsewhere you copy the transcript and paste it yourself.
- Amical's documentation says AI formatting runs the transcript through a language model that adds punctuation and structure.
- Amical's homepage says it can be customized to recognize specific terminology, jargon and proper nouns.
- Amical's pricing page says the free plan includes unlimited local dictation.
- Amical's mobile app privacy policy says audio recordings are not stored on the maker's servers.
- Amical's pricing page lists 'No model training on your data' in its plan features.
- Amical's documentation says it can run entirely on the user's machine, with no data leaving the computer.
- Amical's documentation says dictated words land in whatever app the user has focused.
- Amical's FAQ says the app works offline once a local Whisper model has been downloaded.
- Apple says that, on a Mac, voice inputs and transcripts for general text Dictation are processed on your device and not sent to Siri servers, and that you can check this in Keyboard settings.
- Apple says that with Dictation on a Mac you can speak to enter text in any place where you can type.
- Apple says that once Voice Control is on, you can speak commands to navigate the desktop and apps, interact with what is on the screen, and dictate and edit text.
- Aqua Voice's free plan includes 1,000 free words and its Avalon transcription model.
- Audionotes' homepage describes the app as an AI note-taker that turns voice, audio, video and text into structured notes.
- Audionotes' FAQ says recordings and transcripts are encrypted and stored in the user's account.
- Audionotes' pricing page says the core features are free with unlimited notes and no credit card.
- Audionotes' homepage says user content is not used to train large language models.
- Audionotes' privacy policy names Deepgram and Assembly AI as providers for speech-to-text transcription.
- Audionotes' privacy policy names OpenAI and Google Gemini for AI-based processing and summarization.
- AudioPen's pricing page says its Free plan costs $0, needs no credit card, and allows recordings up to 3 minutes.
- AudioPen's privacy policy says in local mode the audio is transcribed on the device and is never sent to AudioPen or any third party.
- AudioPen's FAQ says all of your data is never used to train any AI.
- AudioPen's FAQ says its iPhone, Android, Mac and Windows apps have an offline option, where transcription runs on the device.
- AudioPen's pricing page says its free plan transcribes recordings and rewrites them in the writing style you pick.
- AudioPen's privacy policy says it may use third-party AI and transcription providers, such as Google and OpenAI, to process text.
- AudioPen's FAQ says it voice types into any app with a text field, such as Gmail, Slack and Notion.
- AudioPen's privacy policy says audio from successfully processed recordings is deleted from its servers within 48 hours.
- BetterDictation's home page says its dictation software never stores audio recordings or transcripts.
- BetterDictation's home page says its Pro add-on adds an extra AI pass to every dictation.
- BetterDictation's home page says that its Pro add-on uses OpenAI to clean the transcripts and post-process the results.
- BetterDictation's home page says its dictation is powered by OpenAI's Whisper and runs entirely on Apple's Neural Engine, so the speech is recognized on the Mac.
- BetterDictation's home page says it types dictated text into any app, in over 100 languages.
- BetterDictation's home page says that it does not need an internet connection and that it operates fully offline, processing voice-to-text locally on the device.
- BiBi Keyboard's homepage says an LLM refines recognized text with punctuation, formatting and preset prompts.
- BiBi Keyboard's homepage lists unified hotword management in its Pro edition.
- BiBi Keyboard's homepage says its open-source edition is completely free.
- BiBi Keyboard's FAQ says the app does not upload audio itself and keeps no copy of it.
- BiBi Keyboard's FAQ says it does not collect personal data and keeps settings and history on the device.
- The BiBi Keyboard documentation says its local models work without a network connection and keep data private.
- BiBi Keyboard's FAQ says audio and text go to the chosen provider when cloud ASR or AI post-processing is on.
- BiBi Keyboard's homepage says its floating ball gives voice input on any keyboard and inserts text automatically.
- BiBi Keyboard's homepage lists editing text by voice instruction as part of its LLM editing feature.
- BiBi Keyboard's FAQ says offline recognition works with local models after they have been downloaded.
- Braina's download page lists a free Lite edition, and says Lite has no dictation feature.
- Braina's download page says Pro dictation includes AI correction, with up to 99 percent accuracy.
- Braina's download page lists Custom Words and Vocabulary among the Pro features.
- Braina's download page says Pro dictation works in any software or website.
- Braina's download page says Pro dictation works both online and offline, in 90 languages.
- SayCopyPaste's website says it puts your words on the clipboard and does not paste them for you.
- SayCopyPaste's website says you add a name, brand or jargon once in Settings, and later dictations get it right.
- SayCopyPaste's privacy policy says it does not retain the audio on its own bravely.dev servers.
- SayCopyPaste's privacy policy says it does not collect voice recordings, which are never stored by the maker.
- SayCopyPaste's website says you can turn on on-device transcription for when you are offline, with lower quality.
- SayCopyPaste's FAQ describes an optional on-device mode for when a phone or tablet is offline.
- Cleft's pricing page lists a Basic plan at $0, with on-device transcription and 5-minute recordings.
- Cleft's privacy policy says it never trains AI models on user data.
- Cleft's pricing page says its AI turns a ramble into structured notes.
- Cleft's privacy policy says only note text, never audio, goes to OpenAI or Groq for summaries.
- Cleft's privacy policy says transcription happens entirely on the iPhone, iPad or Mac, using Whisper AI.
- Cleft's pricing page says Cleft works offline and needs no Wi-Fi.
- Clovy's product page lists a Free tier at $0, for trying Clovy on real work, with free starter credits.
- Clovy's privacy policy says recordings stay on the device by default, unless the user sends them for model processing.
- Clovy's product page says it does not store user content or train on it.
- Clovy's privacy policy names dictation cleanup among features that send audio and transcripts to its model backend for processing.
- Clovy's privacy policy says it routes AI requests to inference providers such as Venice AI.
- Clovy's privacy policy says transcription sends audio to its own backend for processing.
- Clovy's product page says that after a keyboard shortcut, Clovy turns speech into writing in any app on the computer.
- Codictate's homepage says it adapts to the user's custom vocabulary and expands shortcuts.
- Codictate's homepage describes it as a free dictation tool for Mac and Windows, open source under Apache 2.0.
- Codictate's privacy policy says no data is sent to any server during or after transcription.
- Codictate's privacy policy lists audio recordings among the data it does not collect.
- Codictate's privacy policy says neither of its speech models sends data to OpenAI, NVIDIA or any other third party.
- Codictate's privacy policy says audio recordings are processed entirely on the device, with local speech models.
- Codictate's homepage says that after a press of a button and speech, the text is pasted into any app.
- Codictate's privacy policy says the app works entirely offline once the speech models are downloaded.
- Contextli's privacy policy says it does not store audio recordings after a transcription is produced.
- Contextli's pricing FAQ says new accounts start with a free trial rather than a free plan.
- Contextli's privacy policy says it does not use recordings or transcripts to train AI models.
- Contextli's pricing FAQ says its Lightweight AI formats and polishes text with fast, low-cost models.
- Contextli's privacy policy says audio is processed entirely on the device in its Local Whisper Mode on desktop.
- Contextli's privacy policy says it shares audio with Deepgram and transcribed text with xAI / Grok.
- Contextli's home page says it works in any app the user can type in.
- Contextli's pricing FAQ says its offline modes are unlimited on every paid plan and run entirely on the computer.
- DictaFlow's privacy policy says it does not permanently keep audio recordings on its servers after processing in the ordinary course.
- DictaFlow's home page says users can add their own people, products and specialist terms to a vocabulary.
- DictaFlow's home page says its free plan includes 2,000 words every month.
- DictaFlow's home page says its dictation can add optional AI cleanup.
- DictaFlow's home page says speech becomes text on the device, even without an internet connection.
- DictaFlow's privacy policy names OpenAI, Deepgram and Groq as service providers for transcription and related AI processing.
- DictaFlow's home page says it works in supported text fields in email, documents, messaging apps, browsers and AI tools.
- DictaFlow's home page says its desktop AI Mode can revise selected text from a spoken instruction.
- DictaFlow's home page says unlimited offline dictation is included with its Pro plan.
- Dictanote's pricing page says dictation runs through the browser's speech recognition, so the company never receives or stores audio.
- Dictanote's pricing page offers a free plan with 60 minutes of dictation a day in 50+ languages.
- Dictanote's pricing page lists edit by voice, with undo and deleting the last few words, as a Pro feature.
- Dictanote's pricing page lists that it works offline and syncs across devices, on its plans.
- Dictate Keyboard's README says any word can be added to its dictionary with a long press.
- Dictate Keyboard's README says the app is free and open source.
- Dictate Keyboard's README says it offers custom AI rewording of text.
- Dictate Keyboard's README says that with its on-device model, nothing leaves the phone and no network is needed.
- Dictate Keyboard's README says its AI rewording is powered by models from OpenAI, Anthropic and Google Gemini, which are outside companies.
- Dictate Keyboard's README says dictated words appear as clean, formatted text in any app.
- Dictate Keyboard's README says a freeform voice command lets the user speak a request, such as an edit, and the AI carries it out on the selected text.
- Dictate Keyboard's README says it can dictate completely offline with a downloadable model, with no internet connection needed.
- Dictation Daddy's pricing page lists Templates and Custom Vocabulary as a feature of its Solo plan.
- Dictation Daddy's privacy policy says audio recordings are stored on its servers after transcription by default.
- Dictation Daddy's privacy policy says it never uses audio data to train AI models.
- Dictation Daddy's pricing page lists AI Cleanup and Enhancement as a feature of its Solo plan.
- Dictation Daddy's home page says you press one key and dictate into whichever app has the cursor.
- Dictato's FAQ says that a dictation never leaves the Mac.
- Dictato's FAQ says names and jargon are fixed by adding them to the Vocabulary section in Settings.
- Dictato's home page says the full app is free for 7 days, then costs 19.99 euros once.
- Dictato's FAQ says tone rewrites run on Gemma, a model it downloads once and keeps on the Mac.
- Dictato's FAQ says the audio goes to a model running on the Mac, and the finished text lands at the cursor.
- Dictato's home page says nothing is uploaded, even for the cleanup step.
- Dictato's FAQ says that if you can paste into an app, Dictato works there, and lists Gmail, Slack and Word among them.
- Dictato's home page says it keeps working offline, forever, on the Mac it is licensed to.
- Dun's privacy policy says optional writing-style polish uses Apple's on-device Foundation Models when supported and available.
- Dun's privacy policy says personal dictionary entries are used locally to improve recognition and text correction.
- Dun's privacy policy says it does not use your speech to train models.
- Dun's privacy policy says it does not upload your voice for cloud transcription.
- Dun's privacy policy says dictation audio is transcribed on the iPhone using downloaded speech models.
- Dun's home page says that if an app has a text field, Dun can fill it.
- Sayboard's README says it does not send any of your information over the internet, so it keeps no audio on servers of its own.
- Sayboard's README says it does not send any of your information over the internet, so it sends no text to a third-party AI.
- Sayboard's README says it does not send any of your information over the internet, so recordings do not reach the maker's models.
- Sayboard's README describes it as an on-device voice keyboard that uses the Vosk library.
- EnviousWispr's home page says dictation audio never leaves the Mac.
- EnviousWispr's home page says that with a cloud provider, text goes directly to that provider under your own key.
- EnviousWispr's home page says names, jargon and product names you add get fixed on the Mac, before any model reads the text.
- EnviousWispr's home page says the Mac app is free and open source, with no account and no subscription.
- EnviousWispr's home page says your chosen AI model tidies the cleaned text, locally or through your chosen provider.
- EnviousWispr's home page says both speech engines run locally on Apple Silicon, so the audio stays on the Mac.
- EnviousWispr's home page says it turns spoken thoughts into polished text in the apps you use.
- EnviousWispr's home page says your voice stays on your Mac, with offline AI polish.
- Flint's home page says on-device transcription is free forever, while its paid plans cover cloud transcription.
- Flint's privacy policy says recordings are stored on the user's device, not on Flint's servers.
- Flint's home page says it transcribes and summarizes speech into notes, todo lists and key points.
- Flint's privacy policy names Deepgram, OpenAI and Anthropic as the third parties that receive audio or text from cloud features.
- Flint's privacy policy says its fully on-device mode sends no audio or text off the device.
- Flow Keyboard's privacy policy says it does not intentionally store voice recordings on its servers after processing.
- Flow Keyboard's pricing page says registered users can keep using the app for free after the trial, with 5,000 words a week.
- Flow Keyboard's pricing page says AI polishing removes filler words, fixes grammar, adds punctuation and organizes logic, on every plan.
- Flow Keyboard's privacy policy says audio is sent to its servers to be converted to text.
- Flow Keyboard's privacy policy says cloud AI subprocessors process user content to fulfill requests.
- Flow Keyboard's pricing page says it works system-wide on Mac and Windows, and as a keyboard app on iPhone and Android.
- Flow Keyboard's privacy policy describes voice editing: selected text is sent with the user's voice to be edited.
- Flow Keyboard's privacy policy says its voice features need an active internet connection.
- FluidVoice's FAQ says the app is completely free and open source.
- FluidVoice's home page says local processing keeps audio and dictated text on the user's device by default.
- FluidVoice's home page says its Fluid-1 model removes filler, fixes casing and formats text on the device.
- FluidVoice's home page says its Fluid Speech Mini model converts audio to text on the user's device.
- FluidVoice's FAQ says that if the user chooses an optional cloud provider, the content for that feature is sent to it.
- FluidVoice's FAQ says the Mac app needs accessibility permission to type into other apps.
- FluidVoice's home page says its local speech models process dictation on the Mac, with no Wi-Fi needed.
- Yap's product page says audio stays on the user's Mac, so the app does not keep audio on the maker's servers.
- Yap's product page says it is free, open-source dictation for macOS.
- Yap's product page says the user's words land in the field they are using, from the terminal to the browser, so it types into any app.
- Yap's product page says transcription runs on the user's Mac with Apple's Speech framework, so speech is processed on the device.
- Yap's product page says it dictates without a connection once macOS has the speech model for the user's language, so it works offline.
- Yap's GitHub README says it has no network calls, so the app does not send the user's text to a third-party AI.
- FUTO's privacy policy says microphone data will never be transmitted elsewhere, which rules out keeping that audio on FUTO's servers.
- FUTO's privacy policy says microphone data is always processed on the device and will never be transmitted elsewhere.
- FUTO's product page says FUTO Voice Input is designed to work entirely on-device with no data stored.
- Under Gboard's opt-in conventional learning, Google says it will not keep the speech snippets you donate beyond 18 months.
- Gboard's voice typing works in any app where you can type, and Google's example apps are Gmail and Keep.
- The Voice Typing privacy policy says the developer does not receive, store or have access to the audio, so audio is not kept on the maker's servers.
- Voice Typing Notepad's home page says recognition is done by Chrome and the audio goes to Google's speech service, so speech is not processed on the device.
- Voice Typing Notepad's home page says that saying "scratch that" removes the last dictated phrase.
- Ghost Pepper's README lists it as free and open source.
- Ghost Pepper's README says transcription history and recordings are stored on the Mac only, and nothing is sent to the cloud.
- Ghost Pepper's README says a local language model removes filler words and handles self-corrections.
- Ghost Pepper's README says its models run locally on Apple Silicon and nothing is sent anywhere.
- Ghost Pepper's README says you hold Control to talk, then release it to transcribe and paste the text into any text field.
- Glaido's home page says you add names and tech terms once to its dictionary, and it spells them that way from then on.
- Glaido's pricing page says its Basic plan is free forever, with 2,000 words a week.
- Glaido's home page says it offers optional AI layers for formatting, prompts and tool integrations.
- Glaido's privacy policy says it relies on service providers for transcription, formatting and AI features.
- Glaido's privacy policy names Groq, Inc. as a provider for AI processing.
- Glaido's home page says it works everywhere you work, with one hotkey in any app.
- Glaido's Commands documentation says you speak an instruction, and Glaido rewrites, translates or summarizes text for you.
- Glimpse's home page lists a dictionary and replacements among the free features.
- Glimpse's home page says its Free plan is $0 forever, with unlimited dictation in any app.
- Glimpse's privacy page says audio and transcripts stay on the user's machine and never touch its servers.
- Glimpse's home page says a one-time license adds AI cleanup, along with meeting recordings.
- Glimpse's privacy page says that when you use an optional LLM feature, the text and prompt go from your machine to the provider you set up.
- Glimpse's home page says it runs its speech models on the user's own computer.
- Glimpse's home page says it types what you say into whatever you are using.
- Glimpse's home page says you select text, say how to change it, and the text is rewritten in place.
- Glimpse's home page says its dictation works with the wifi off.
- Google says that after you start voice typing in Google Docs, you can use commands to edit and format your document.
- Google says Eloquent uses AI to capture your intended meaning, so it does not transcribe stumbles and filler words verbatim.
- Google says your audio never leaves your iOS device, except that some advanced optional features require the cloud.
- Google says Eloquent is free of charge with no usage limits.
- Google says the Eloquent keyboard for iOS is coming soon, so the keyboard is not yet available.
- Google says Eloquent creates an editable dictionary by learning from your vocabulary to improve the accuracy of dictation.
- Google says all machine learning processing in Eloquent runs locally on the iOS device.
- Google says Eloquent works even when you are offline.
- Google says text you speak with Gboard's advanced voice typing stays on your device, except when you use Fix it or detailed edits.
- The App Store listing for Gboard, published by Google, shows the app as Free.
- Google says you can edit dictated text with voice commands using Gboard's detailed edits.
- Google says you can choose to donate snippets of your speech to improve its speech models.
- Google says you can compose and rewrite text by voice using Gboard's writing tools.
- Google says its voice typing and captions work in Google Docs and in the speaker notes and captions of Google Slides. Its help page for the feature does not list Google Sheets.
- Handy's privacy policy says that when you invoke cloud post-processing, it sends the transcription and prompt to the provider you selected.
- Handy's privacy policy lists custom words among the settings it stores locally on your device.
- Handy's project page says it is free, and that accessibility tooling belongs in everyone's hands, not behind a paywall.
- Handy's privacy policy says that during ordinary transcription it does not send your microphone audio to the maintainers or to Handy servers.
- Handy says it offers optional post-processing of transcripts through AI providers you choose, local services or custom endpoints.
- The Handy project describes itself as a free, open source speech-to-text application that works completely offline.
- Handy's README says that transcription happens on your own computer without sending any information to the cloud.
- Handy's README describes its use as pressing a shortcut, speaking, and having your words appear in any text field.
- The HEX README says Corrections in Modes fix names and preferred spellings before the text is pasted.
- The HEX README says you hold a shortcut, speak and release it, and that HEX transcribes on your Mac and pastes the result into the app you are using.
- HEX describes itself as local-first voice dictation for Apple silicon Macs running macOS 15 or newer.
- The HEX README says that when post-processing with OpenCode is turned on in Modes, the app sends the transcript, the foreground application and the Brave website hostname to the configured model provider.
- The hns documentation says audio is processed entirely on the local machine, using a local Whisper model.
- The hns documentation says transcribed text is copied to the clipboard, ready to be pasted anywhere, so it does not type into other apps.
- The hns documentation says audio is processed locally and that no data leaves your device.
- The hns homepage says your sensitive information and private thoughts are not harvested for model training or advertisement.
- The hns homepage says it has no subscription and no recurring fee for something you can run on your own device.
- The hns documentation says that after the initial model download, no internet connection is required to transcribe.
- Whisper To Input's readme says you can open any app that needs text input, such as a browser, and use the app's keyboard to dictate into its input box.
- Keller's Chrome Web Store listing says the developer will not collect your data, so audio is not kept on the developer's servers.
- Keller's Chrome Web Store listing says the developer will not collect or use your data.
- Keller's Chrome Web Store listing says the Whisper AI engine can work fully offline once it has loaded.
- Keller's Chrome Web Store listing says captured text goes into the extension's own text area, and you copy it from there with Ctrl+C.
- Keller's Chrome Web Store listing says the Whisper AI engine runs locally in your browser.
- KeyVox's home page calls its voice transcription app always free.
- KeyVox's home page says its custom dictionary holds names, emails and special words.
- KeyVox's privacy policy says recordings and transcripts are not uploaded to its servers for transcription.
- KeyVox's FAQ says no audio or text ever leaves the Mac.
- KeyVox's FAQ says no internet connection is required on the Mac after the model is downloaded.
- KeyVox's home page says all dictation and transcription happens on the device.
- KeyVox's FAQ says the Mac app needs Accessibility permission to insert text into other apps.
- KeyVox's home page says Vibes gives on-device, reversible writing styles for anything you say.
- Koe's documentation says Apple Speech is local on macOS 26 and later, and the MLX and sherpa-onnx options need the Koe MLX build.
- Koe's documentation says it writes to the clipboard and auto-pastes when Accessibility permission is available.
- Koe's sponsor page says Koe is completely free to use and that there is no current plan to charge for it.
- Koe's documentation says an optional LLM profile corrects the text through OpenAI, Anthropic, a compatible gateway, APFEL or local MLX.
- Koe's privacy policy says the Koe MLX build can run speech recognition and the LLM locally after the model weights are downloaded.
- Koe's privacy policy says requests go from the Mac to the provider in the config, not through a Koe-operated server.
- Koe's home page says its dictionary entries are plain text that can be generated, reviewed and edited in bulk.
- Koe's privacy policy says remote LLM providers receive the rendered prompts, transcript, revision context and selected dictionary entries.
- LazyTyper says it is currently free to download and use, with no ads.
- LazyTyper says it does not store or collect voice recordings or audio data.
- LazyTyper says 5 of its speech models run entirely on your device.
- LazyTyper says 5 of its 12 speech models are offline local models that run entirely on your device.
- Letterly's home page says you can dictate in any app and gives Gmail, Slack and Notion as examples.
- Letterly's plans page lists a Dictionary as a feature on both the Free and Pro plans.
- Letterly's help center says recordings, notes and rewritten text are never used to train AI models.
- Letterly's help center says recording works offline, but transcription and rewriting need the internet.
- Letterly's help center says recordings and notes go to its servers over an encrypted connection and are stored on Google Cloud.
- Letterly's help center says a language model rewrites a transcript into the format the user needs.
- Letterly's help center says it sends audio and text to third-party providers so they can transcribe or rewrite it.
- Letterly's help center says transcription and rewriting run on AI models in the cloud, so they need the internet.
- Letterly's help center says Letterly has a Free plan, alongside its Pro plan.
- When you enable Dictation, choosing Share Audio Recordings allows Apple to store audio of your Siri and Dictation interactions, and Apple may review a sample of it; choosing Not Now declines.
- MacParakeet's 0.8.10 release notes describe AI cleanup of a long transcript.
- MacParakeet says your audio never leaves your Mac, and all transcription runs locally on your device.
- MacParakeet lists custom vocabulary among its features.
- MacParakeet says it is free and open-source under the GPL-3.0 license, with no paid tier.
- MacParakeet says its core speech recognition runs on the Apple Silicon chip in your Mac, with no cloud speech servers.
- When you use MacParakeet's optional AI summaries or chat, only the transcript text is sent to a provider you choose.
- MacParakeet says it can type into any app that has a text cursor on the Mac.
- MacParakeet says that after the speech model is downloaded during setup, dictation and file transcription work completely offline.
- MacWhisper's homepage says its dictation lets users use their voice in any app.
- MacWhisper's own website lists its Free plan at โฌ0 per month, free forever.
- MacWhisper says it processes sensitive content locally, without data ever leaving the Mac.
- MacWhisper's privacy policy says that no audio, text or other data leaves the user's device.
- MacWhisper's homepage lists Offline as a feature, and says it processes sensitive content locally.
- MacWhisper's homepage says users can switch between AI models hosted remotely or locally on their Mac.
- Whisper Memos's privacy policy says it uses the GPT-4 API to structure transcribed text into organized paragraphs.
- Whisper Memos's privacy policy says it sends transcribed text to the GPT-4 API to be structured into paragraphs.
- Whisper Memos's privacy policy says transcription sends each audio file to OpenAI, so the speech is not processed on the device.
- Microsoft's SwiftKey help page says offline recognition for its Google Voice Typing mode is set up in the Google app.
- Microsoft says SwiftKey may send anonymized voice clips to Microsoft, so speech is not handled only on the device.
- Microsoft says SwiftKey may send anonymized voice clips to Microsoft to improve recognition, and that users can opt out in SwiftKey settings.
- Microsoft says you can enter text on your PC by speaking with voice typing.
- Windows voice typing needs an internet connection, a working microphone and the cursor in a text box before it will type what you say.
- Microsoft says it keeps voice clips that people contribute for sampling and listening for up to two years.
- Microsoft says voice clips will occasionally be sampled and listened to for training its speech recognition models.
- Microsoft says Windows voice typing uses online speech recognition, and that the online recognition is powered by Azure Speech services.
- Monologue's home page says its AI providers process dictation, with zero data retention.
- Monologue's home page says that for Dictation it does not save audio or transcripts on its servers.
- Monologue's FAQ says it offers local transcription models on Mac that turn speech into text without an internet connection.
- Monologue's privacy policy lists training any artificial intelligence models on users' audio or transcripts among the things it will not do.
- Monologue's home page says it has a dictionary for names, acronyms and company vernacular, used everywhere.
- Monologue's home page says dictation gives clean, formatted text in any app.
- Monologue's home page says its local models work with no connection required.
- On macOS Monterey 12 with Apple silicon, general text dictation was processed on the Mac in many languages, and no internet connection was required.
- Muesli says OpenAI, OpenRouter or ChatGPT may receive meeting transcript text for optional AI meeting summaries.
- Muesli describes itself as free dictation and meeting transcription for the Mac.
- Muesli says it pastes dictated text into the app the user is already using.
- Muesli's privacy policy lists personal dictionary entries among the data it processes locally.
- Muesli says the default speech-to-text path does not send recordings to Muesli servers.
- Muesli says its core speech transcription runs locally on your device after a model is installed.
- Muesli says it works without internet after install.
- Murmure's website says you can bring your own local or remote LLM to process your text.
- Murmure's documentation says a spoken instruction is applied to the selected text by the language model, which then replaces the selection.
- Murmure's documentation says its dictionary helps the app recognize words it might otherwise miss or misspell.
- Murmure's website says the app is free and open-source.
- Murmure's documentation says the transcribed text is inserted into the focused application.
- Murmure's privacy policy says it does not upload recordings or transcriptions to a cloud service by default.
- Murmure's privacy policy says that if the configured LLM address is a remote server, the transcription data is sent to that server.
- Murmure, an open-source Windows dictation app, says it turns speech into text with no internet connection and zero data collection.
- Murmure's README says it transcribes speech locally with the Parakeet TDT 0.6B v3 model.
- The nerd-dictation README says a JSON grammar file restricts the phrases that VOSK recognizes, for better accuracy.
- The nerd-dictation README says its output can simulate keystroke events, which is the default.
- The nerd-dictation project describes itself as offline speech to text for desktop Linux, built on the VOSK-API.
- The nerd-dictation README says the speech model is a local folder, by default in the user's configuration directory.
- The README of Offline Voice Input says no audio ever leaves the phone.
- The README of Offline Voice Input says speech is transcribed entirely on the device.
- The README of Offline Voice Input says the transcribed text is typed into any app.
- The README of Offline Voice Input says the app transcribes speech with no network connection required.
- Nuance's Dragon help says users can dictate text into most Windows applications in three steps.
- Nuance's Dragon Professional 16 help says it describes how to add words and phrases to the vocabulary.
- Nuance's Dragon Professional 16 help lists the spoken commands that delete the last text dictated.
- Nuance's privacy statement says it may capture users' voice, and that voice snippets may be reviewed to build, train and improve its automated methods.
- Openbroca's privacy policy says audio and text go to AI model providers, which handle them under their own terms.
- Openbroca's pricing page says its text refinement turns natural speech into clear text with useful structure.
- Openbroca's privacy policy says it does not store your audio.
- Openbroca's pricing page says the personal dictionary lets you add preferred words and learns corrections after dictation.
- Openbroca's pricing page says the app has a free plan of 3,000 text units each week.
- Openbroca's home page says its speech recognition and text cleanup run in the cloud and need an internet connection.
- Openbroca's privacy policy says the app sends your audio to a third-party AI provider for speech recognition.
- Openbroca's pricing page says it inserts text into the focused field, with a copy fallback when needed.
- OpenLess's README says its text is AI-polished as it streams to the cursor.
- OpenLess's README says all of your data stays on your machine.
- OpenLess's README says proper nouns go into its dictionary as hotwords.
- OpenLess's README says it inserts the result at the cursor.
- OpenLess's README says it bundles the Qwen3-ASR local speech model for macOS.
- OpenLess's README lists OpenAI and Google Gemini among its polish providers.
- Open Planet Software's privacy policy says Just Press Record audio and transcripts are never uploaded to the company's servers.
- Open Planet Software's privacy policy says Just Press Record typically converts audio to text on the iOS device, though audio may go to Apple's iCloud servers.
- OpenSuperWhisper's README lists a custom dictionary and keyword boosting on its contribution to-do list.
- OpenTypeless says its AI polishing is optional and fixes grammar, punctuation, tone and structure.
- OpenTypeless says that with BYOK it sends transcript text to the LLM endpoint the user configures.
- OpenTypeless's home page says its custom dictionary takes names, product terms and technical vocabulary, and can be searched, edited, imported and exported.
- OpenTypeless's pricing page says the app is free forever under the MIT open-source license.
- OpenTypeless's home page says users speak naturally in any app, with optional AI polish of the text.
- OpenWhispr's dictation cleanup page says that by default it runs dictation through a language model, which rewrites the text before it is pasted.
- OpenWhispr's cleanup page lists Cloud Providers as an option, using the user's own key with OpenAI, Anthropic, Gemini or Groq.
- OpenWhispr's privacy policy says words added to the custom dictionary are stored on the device and sent to the transcription and cleanup provider as a hint.
- OpenWhispr's product page says you can download the app and dictate into any app on your computer.
- OpenWhispr's product page says it is free to start on macOS, Windows and Linux, with no card and no account required.
- OpenWhispr's privacy policy says audio sent to OpenWhispr Cloud is processed in real time and not stored by OpenWhispr.
- OpenWhispr's privacy policy says local transcription runs fully on the device and that no audio is transmitted in that mode.
- OpenWhispr's privacy policy says it never uses transcribed content, notes or audio to train its own AI models.
- OpenWhispr's README lists Whisper and NVIDIA Parakeet among the local speech models that run offline.
- OSTT's home page says it can run AI prompts or shell commands on a transcription.
- OSTT's command reference says its keywords improve transcription accuracy for technical terms, names and domain vocabulary.
- OSTT's providers reference says local models run on the user's machine and need no API key.
- OSTT's getting-started guide says it can transcribe offline with downloaded local models.
- OSTT's home page says it is open source and has no subscription.
- OSTT's home page says it pastes the transcription into any focused app.
- OSTT's why-OSTT page says it stores recordings locally so the same audio can be transcribed again.
- Otter's privacy policy says it discloses personal information to artificial intelligence service providers that give backend support for some of its product features.
- Otter's pricing page lists custom vocabulary for names, jargon and acronyms, with a limit of 5 terms on the Basic plan.
- Otter's security page says it uses AWS S3 storage for data, with server-side encryption on the S3 buckets.
- Otter's privacy policy, effective June 16, 2026, lists training its proprietary AI technology on de-identified audio recordings and on transcriptions among the purposes for which it uses information.
- Paraspeech's home page says it polishes text with Cloud Cleanup or, where eligible, local rewriting in the direct-download Mac app.
- Paraspeech's privacy policy says it stores feedback audio that users choose to send, deleting recordings after 30 days.
- Paraspeech's privacy policy says it may send transcript text and rewrite instructions to Groq for cloud rewriting.
- Paraspeech's home page says it inserts text into many editable Mac fields and needs manual paste where insertion is unsupported.
- Paraspeech's home page says local speech models transcribe the recording directly on the user's Mac.
- Paraspeech's pricing page offers a 7-day trial, after which the subscription is billed.
- Paraspeech's documentation says its word replacement rules fix names, terms or phrases that are repeatedly misheard.
- Paraspeech's home page says that once a local model is downloaded, it works without an internet connection.
- Pithflow's home page says its AI cleanup removes filler words, fixes punctuation and applies a chosen tone.
- Pithflow's privacy policy says it does not store audio after transcription completes.
- Pithflow's privacy policy describes a dictionary of terms with a 25-term limit on the free plan.
- Pithflow's pricing page lists a free plan at $0 with 2,000 words a week.
- Pithflow's home page says transcribing and cleaning a dictation happens on its servers, not on the device.
- Pithflow's privacy policy says the transcribed text goes to Groq for cleanup by a language model that Groq runs.
- Pithflow's home page says it types directly into the focused text field of any app.
- Quill Flow's home page says its AI handles punctuation, capitalization and grammar automatically.
- Quill Flow's home page says its personal dictionary holds your names, brands and shortcuts.
- Quill Flow's FAQ says the app is free to download, with premium features at $3 or $4.99 a month.
- Quill Flow's privacy policy says voice recordings are discarded once each request is complete and that nothing is saved on its servers.
- Quill Flow's FAQ says it works in every app where you can type, because it is a keyboard extension.
- Rekody's developer docs say an optional LLM pass cleans filler words, fixes grammar and adapts formatting to the app.
- Rekody's dictionary help page says it teaches the app names, jargon and product terms.
- Rekody's home page describes the app as free and open source for the Mac.
- Rekody's privacy page says it runs no servers that could receive audio.
- Rekody's help page on reviewing dictation data says Rekody does not train a model for the user.
- Rekody's home page says transcription happens on the user's Mac by default.
- Rekody's developer docs say only the transcript text is sent to an LLM provider, never audio.
- Rekody's home page says the text lands at the cursor in any app.
- Rekody's privacy page says its speech model runs fully offline after a one-time download.
- Samsung's keyboard settings have an Offline voice input section where you install offline language packs.
- The WhisperWriter README lists using GPT for post-processing under its roadmap, so the app does not yet rewrite or format text with AI.
- The WhisperWriter README says the app uses a local model by default, so speech is transcribed on the user's computer unless they switch to the OpenAI API.
- The WhisperWriter README says the transcribed text is written automatically to the active window, so it types into the app the user is working in.
- Sayso's website says a cloud model or an AI style is your choice, with your own key.
- Sayso's website says a cloud model or an AI style is your choice, with your own key, and that the app names each place your words go.
- Sayso's website says the app is free and you pay what you want.
- Sayso's website says your voice does not leave your computer, so recordings do not reach the maker's models.
- Sayso's website says a speech model on your own computer does the transcription.
- Sayso's website says the dictated text lands in the app you are using.
- Sayso's website says your voice does not leave your computer.
- Speech Recognition Anywhere's own page describes only a limited 30-day free trial, with paid subscriptions after it.
- Speech Recognition Anywhere's page documents a built-in undo command that reverses the last text command.
- Speech Recognition Anywhere's page says an optional Windows Add-On lets speech recognition work with Microsoft Windows programs such as Office and Word.
- Sotto's home page says dictation can be processed by AI before it is pasted, using a function picked from the recording bar.
- Sotto's privacy policy says that by default audio and transcripts are not sent to Sotto.
- Sotto's privacy policy says text processed by a cloud AI provider is sent directly from the Mac to that provider.
- Sotto's home page says users can add their own terms, names and jargon to a dictionary that is used as hints for transcription.
- Sotto's privacy policy says its built-in Whisper and Parakeet models run entirely on the user's Mac.
- Sotto's home page says transcripts paste directly into any app, with no copy and paste.
- Sotto's FAQ says dictation and transcription work with no internet connection after a one-time model download.
- SpeakApp AI's privacy policy says AI service providers process the data it sends them on its documented instructions.
- SpeakApp AI's privacy policy says hosting providers store users' audio recordings and transcriptions in EU data centers.
- SpeakApp AI's home page says it can change the tone of text and rewrite it with AI, for example by applying an email filter.
- Speakly's FAQ says AI removes filler words, fixes mistakes and formats the dictated text.
- Speakly's FAQ says the app is free to download and use, with 4,000 words per week on the free plan.
- Speakly's privacy policy says queries may be sent to the OpenAI or Anthropic API to serve requests.
- Speakly's home page says it types the polished text where the cursor is, in any app.
- SpeakoFlow's AI cleanup runs a language model over dictated text to fix grammar and punctuation, and it is off by default.
- SpeakoFlow's privacy page says the assistant sends text only to the provider the user chose, which can be a cloud AI provider with the user's own key.
- SpeakoFlow's custom words page says users can teach the app the names and terms it gets wrong.
- SpeakoFlow's privacy page says the feedback reports sent to the developer include no recordings or transcripts.
- SpeakoFlow's home page FAQ says the app has no paid plan, no trial, no word cap and no account.
- SpeakoFlow's privacy page says speech is transcribed on the user's computer unless the user picks a cloud speech service.
- SpeakoFlow's AI cleanup follows the user's spoken edits, which the page describes as part of the cleanup step.
- SpeakoFlow's dictation guide says it types spoken words into any app where the cursor can be placed.
- SpeakoFlow's home page FAQ says dictation works offline, running a speech model on the user's computer.
- SpeakType's home page says the app is free and open source, with no subscriptions and no limits.
- SpeakType's home page says nothing leaves the device after on-device transcription.
- SpeakType's home page says its AI transcribes speech on the device, entirely on the Apple Silicon chip.
- SpeakType's home page says spoken words appear in any text field, in any app.
- SpeakType's home page describes its macOS dictation as offline transcription on Apple Silicon.
- SpeakType's home page says no data ever leaves the Mac, with no cloud service, accounts or tracking.
- The Speech Note README says no data is sent to the Internet.
- The Speech Note README says no data is sent to the Internet, so text is not sent to a third-party AI service.
- The Speech Note README says voice processing takes place locally on the user's computer.
- The Speech Note README says a global keyboard shortcut can insert decoded speech into any focused window.
- The Speech Note README says text and voice processing run entirely offline on the user's computer, without a network connection.
- Speechnotes lists its Dictation Free plan at $0 per month, with an online dictation notepad and a voice typing Chrome extension.
- SpeechTexter's home page calls the app free and says no download, installation or registration is required.
- SpeechTexter's privacy policy says all speech is processed at Google servers, which carry their own privacy policy.
- SpeechTexter's home page says the app has a customizable voice commands list that includes undo, redo and making a new paragraph.
- Spokenly's privacy policy says that when a Mode uses AI Instructions, the transcription and any optional context may be sent to AI services, even if transcription runs locally.
- Spokenly's home page says that pressing a shortcut and speaking makes the words appear at the cursor in any application.
- Spokenly's documentation says custom dictionaries help transcription models recognize names, brands and technical terms, through a Pro subscription or with the user's own API key.
- Spokenly's Free plan costs nothing, with unlimited use of Whisper and Parakeet models and bring-your-own API keys.
- Spokenly's privacy policy says that with local models, audio recordings and transcriptions never leave the device and all speech-to-text processing happens on the device.
- Spokenly's pricing page says its free plan works completely offline, with unlimited use of the Whisper and Parakeet models.
- Spokenly's privacy policy says its backend temporarily keeps an encoded copy of the audio for its ChatGPT and Claude plugins, and that each copy expires no later than one hour after it is stored.
- Spokenly says its optional on-device cleanup model removes filler words and repeats, applies corrections the speaker says out loud, fixes punctuation and capitals, and writes numbers as digits.
- The Talk-type README lists DeepSeek, Anthropic, OpenAI, xAI, Qwen and Kimi as the AI providers its text cleanup supports.
- The Talk-type README says AI cleanup is optional, and that the user picks a provider and enters its API key and model ID.
- The Talk-type README says it writes into the focused input field through macOS Accessibility, or keeps the text on the clipboard when that is unavailable.
- TalkaType's home page offers a free Chrome extension, with a button to get it.
- TalkaType's home page describes it as a browser feature that lets you input text on any website using your voice.
- TalkaType's support page says the network communication needed to complete recognition is required.
- Talon's license agreement grants a royalty-free license to install and use the software.
- Talon's home page demonstrates spoken commands for editing text, such as select all and delete word.
- Typeless's privacy policy says your voice audio and context are processed in real time on its cloud servers, not on the device.
- Typeless's privacy policy says it may send your audio and relevant text to third-party cloud AI providers, including Microsoft and Google.
- Typeless's home page says its dictation data is never used for model training by it or any third party.
- Typeless's pricing page says a new account starts on its Free plan, which has limited words each week.
- Typeless's privacy policy says no voice recordings or screen context data are stored on its servers.
- Typeless's home page says its personal dictionary recognizes the names, terms and expressions that matter to each user.
- Typeless says that when you speak, it turns your words into polished messages, emails and documents.
- Typeless's home page says you can select text and speak a command to make it shorter, longer or change its tone.
- Typeless says it works across all the apps you use on Mac, Windows, Linux, iOS and Android.
- TypeWhisper says its workflows remove filler words, fix punctuation, format notes and translate, using the AI provider the user chooses.
- TypeWhisper's privacy policy says audio processing, transcription and storage happen on the user's device by default.
- TypeWhisper says cloud speech and AI providers are used only after the user configures and selects them.
- TypeWhisper's macOS documentation says its Dictionary lets the user add custom terms and define corrections for commonly misrecognized words.
- TypeWhisper's pricing page says its free GPLv3 tier is free to install and run, including internal use.
- TypeWhisper's macOS documentation says that in dictation-only workflows a spoken instruction is detected, removed from the text and applied to it.
- TypeWhisper's privacy policy says the optional correction contributions exclude audio, so recordings are not part of that path.
- TypeWhisper says its local speech engines run on the user's device.
- TypeWhisper says it types dictated text into the active text field in any app, with rules per app or website.
- TypeWhisper says local transcription can function offline once local models are downloaded.
- Utter's home page says that when a user chooses a mode for email, notes, summaries or chat, Utter rewrites the words with the right structure and tone.
- Utter's privacy policy says it does not store audio recordings or transcript text on its servers after processing is complete.
- Utter's home page says the user can add replacements for names, jargon and phrases so Utter gets their wording right.
- Utter's home page says the free plan lets the user use local models or their own API keys without an Utter subscription.
- Utter's privacy policy says voice and transcript content is not used to train its models or create behavioral profiles.
- Utter's privacy policy says transcript text may be sent to the AI provider chosen in the settings when optional AI features are enabled.
- Utter's offline guide says transcription with a downloaded local model runs on the device.
- Utter's home page says it works with any app where the user can type text.
- Utter's home page says it turns speech into text on the device, so dictation works without a connection.
- Vocalinux's privacy policy says local engines do not send the user's voice to the maker's servers.
- Vocalinux's home page says the app is free and open source under AGPL-3.0, with no premium tiers.
- Vocalinux's privacy policy says the installed app does not send voice recordings or audio data to the maker.
- Vocalinux's privacy policy says local engines do not need the internet after the model is downloaded.
- After a model is downloaded, Vocalinux runs speech to text on your machine with whisper.cpp by default, and it also supports OpenAI Whisper, Faster Whisper, VOSK and Parakeet.
- Vocalinux turns speech into typed text in whatever app has focus, which makes it a dictation tool rather than a voice command tool.
- Voibe's privacy page says audio is never stored and never used to train AI in its zero-retention cloud mode.
- Voibe says that in Cloud Mode an open-weight language model, GPT-OSS 120B, cleans up the transcribed text.
- Voibe's pricing page includes custom vocabulary in the Voibe Dictation plan, and the free plan also lists it.
- Voibe's pricing page says Voibe Free costs nothing and never expires, with dictation in any app limited by words.
- Voibe's privacy page says Voibe never stores a word of the dictated voice.
- Voibe says On-Device Mode works with no internet connection at all, while Cloud Mode needs one.
- Voibe says that in On-Device Mode your voice is transcribed entirely on your Mac.
- Voibe's home page says your voice is typed into any app.
- Voice Control on a Mac accepts up to 1,000 custom vocabulary terms, and you can record how each is pronounced.
- Voice In's website lists a free plan that supports voice typing on popular sites such as Gmail and Google Docs.
- Voice In's website says no audio or transcript is ever sent to the developer's servers.
- The Voice In privacy policy says the extension does not transmit dictated audio or transcribed text to external servers.
- Voice In's website says its advanced features include custom voice commands to edit and automate text.
- Voice Control for ChatGPT's home page offers a free extension for Windows and macOS.
- Voice Control for ChatGPT's home page says it records your voice and submits your spoken questions to ChatGPT, a third-party AI.
- Voice Control for ChatGPT's support page says that completing speech recognition needs network communication, so the extension is not described as working offline.
- VoiceInk's privacy policy says that with optional AI enhancement, only the transcribed text, never the audio, is sent to the provider the user selects.
- VoiceInk's website says you can use your voice to write anywhere, in any application.
- VoiceInk's privacy policy says it does not store app transcriptions or audio files on its servers.
- VoiceInk's website says you can teach it the words you use, so it preserves names, product names and unusual spellings.
- VoiceInk's FAQ says it processes all voice transcription locally on your device, and that your voice data never leaves your Mac.
- VoiceInk's Modes documentation says the Enhancement mode cleans up dictated text with AI, using selected text and screen context when available.
- VoiceInk's website says you select a passage and say how it should change, and the selected text is replaced in place.
- The VoiceInk README says the latest version is available with a free trial from tryvoiceink.com.
- VoiceInk's README claims that 100% offline processing keeps the user's data on the device.
- Voice Notebook's help page says its speech recognition runs on Google back-end servers, not on the device.
- Voice Notebook's home page says that installing the Windows, Mac or Linux integration module lets you use voice to input text in any application on those systems.
- Voice Notebook's help page says the service does not work offline, because it uses Google back-end servers for speech recognition.
- Voicetypr's home page says a user can connect their own OpenAI, Anthropic or Gemini key to turn rough dictation into clean prompts, emails and notes.
- Voicetypr's home page lists custom vocabulary among its dictation features, next to push-to-talk and translation.
- Voicetypr's privacy policy says that with a local model selected, its hosted API does not receive the raw microphone recording.
- Voicetypr's privacy policy says the transcribed text and formatting instructions go to the AI provider or endpoint the user configures.
- Voicetypr's privacy policy says that with a downloaded Whisper or Parakeet model, recorded audio is processed on the user's computer.
- Voicetypr's FAQ says it pastes dictated text into apps such as Gmail, Slack, Notion, Word and Cursor, anywhere the user types.
- Voicetypr's FAQ says its dictation is offline by default, so transcription happens on the user's machine.
- VoiceVibe's pricing page lists AI text transformation, which edits or creates content, changes tone, fixes grammar and translates, as a Pro feature.
- VoiceVibe's pricing page lists a personal dictionary and context for improved accuracy as a Pro feature.
- VoiceVibe's pricing page lists a Free Forever plan at $0 that covers Voice Typing mode only.
- VoiceVibe's privacy policy says it keeps each voice clone's audio, transcript and derived model data until the user deletes it or 1,095 days pass without use.
- VoiceVibe's privacy policy says it does not use user data to train AI models.
- VoiceVibe's privacy policy says the relevant conversation context is sent to VoiceVibe and its AI providers for processing.
- VoiceVibe's privacy policy says that on the free tier, voice audio and dictated text are processed on the device and stay on the computer.
- VoiceVibe's home page says that when the user holds the key and speaks, the dictation is typed into any text field.
- VoiceVibe's home page says the user can select text, hold the key and say commands such as 'shorter', 'more formal' or 'translate to Spanish'.
- VoiceVibe's privacy policy says free voice typing keeps voice audio and dictated text on the user's computer, and works offline.
- Voicy's home page lists AI rewriting for professional emails, with automatic punctuation and formatting.
- Voicy's privacy policy says it does not store any recording or transcription data.
- Voicy's FAQ says it does not use recordings to train an AI model or for any other purpose.
- Voicy's privacy policy says its transcription and post-processing steps are done through Groq.com.
- Voicy's privacy policy says that in offline dictation on its Mac, Windows and Linux apps, audio is transcribed on the device and not sent to Voicy's servers or Groq.
- Voicy's home page says it works on all the websites and apps the user uses daily.
- Voicy's FAQ says offline mode on Mac, Windows and Linux transcribes on the computer, with nothing sent to its servers.
- The VOXD README says it offers AI rewriting of transcripts, with local or cloud language models.
- The VOXD README says the software is open source and free of charge.
- The VOXD README lists OpenAI, Anthropic and xAI as supported providers for its AI post-processing.
- The VOXD README says its voice processing runs locally, on the machine.
- The VOXD README says it provides voice typing in any Linux app.
- The VOXD README says its voice processing is local and offline.
- The Voxtype home page says it offers an optional post-processing pipe through any LLM or shell script.
- The Voxtype configuration reference says its initial prompt gives Whisper context for domain-specific vocabulary.
- The Voxtype home page says the audio stays on the user's machine, with no cloud and no telemetry.
- The Voxtype FAQ says speech recognition runs on the machine with local whisper.cpp models.
- The Voxtype FAQ says the transcribed speech is typed at the cursor position or copied to the clipboard.
- The Voxtype FAQ says all speech recognition is done locally with whisper.cpp.
- The Weesper Neon Flow pricing page says the app uses custom dictionaries for legal and other specialist terms.
- The Weesper Neon Flow home page says users try all features for 15 days, then choose a plan.
- The Weesper Neon Flow privacy policy says audio is not sent to the maker's servers for transcription.
- The Weesper Neon Flow privacy policy says speech recognition runs locally, offline, on the user's device.
- The Weesper Neon Flow home page says clean text appears at the cursor in Gmail, Slack, Notion, Word and other desktop apps.
- The Weesper Neon Flow home page says it offers 100% offline voice dictation for Mac and Windows.
- Whisper To Input does its recognition on a backend, either the OpenAI API with your own API key or a Whisper server you host yourself.
- Whispering's repository says the app is free and open source.
- Whispering's README says dictating into other apps is not part of the current build.
- Whispering's README says it has no on-device transcription models, so speech is sent to a transcription endpoint.
- Whispering's README says it has no on-device transcription models and that transcription needs an endpoint.
- Whispering's README says it has no accounts, sync or hosted transport, and that its recordings are local.
- Whispering's README says its Text connection is a base URL and an optional API key that the user sets.
- Whispering's README says it turns a transcript into another text with prompts the user writes.
- WhisperTyping's AI modes page says its Rewrite mode rewrites the user's selected text based on their instructions.
- WhisperTyping's privacy policy says it never stores audio or transcripts, so it does not keep audio on its servers.
- WhisperTyping's pricing page lists custom vocabulary among the features of its free plan.
- WhisperTyping's home page says a free plan is available with no credit card required.
- WhisperTyping's privacy policy says the text of its AI features is sent to a third-party AI provider, OpenAI, for processing.
- WhisperTyping's privacy policy says it routes audio to cloud-based speech-to-text providers, so its speech recognition is not done on the user's device.
- WhisperTyping's FAQ says the transcribed text is automatically typed through a virtual keyboard, so it works in any application.
- WhisperTyping's FAQ says a user can say 'Rewrite this more formally' to edit selected text, so it takes voice commands for editing.
- WhisperTyping's FAQ says it requires an internet connection because its speech recognition is processed in the cloud, so it does not work offline.
- Willow's pricing page lists 'Add dictionary terms' among its dictation features.
- Willow's security page says customer content is processed by Willow or its applicable AI subprocessors, without retention for enterprise teams.
- Willow's privacy policy says Willow and its service providers do not retain voice audio or use it for model training, in Private Mode.
- Willow Voice's Basic plan is free, with no credit card needed, and includes an unlimited weaker speech-to-text model.
- Willow's security page says no audio is stored after transcription, and that the recording is discarded.
- Willow's security page says Willow is offline when you need it.
- Willow's site describes Scribe as AI writing from your intent.
- Willow's pricing page lists 'Works anywhere you type' among the features of its Basic plan.
- Microsoft's English (United States) voice typing table lists 'Delete that', 'Erase that' and 'Scratch that' for deleting the last spoken word or phrase, and 'Select that' for selecting it.
- Microsoft says fluid dictation in voice typing is powered by small language models that run on the device, for fast and private processing.
- Wispr Flow's pricing page says the free plan includes speech to text in any app in 100+ languages.
- Wispr Flow's pricing page lists its Free plan at $0 per month, including speech to text in any app and 100+ languages.
- Whisper IME's readme says voice recognition works entirely offline once the models are downloaded, so recordings are not sent to a server for recognition.
- Whisper IME's readme says voice recognition works entirely offline once the models are downloaded, so the recognition runs without a network connection.
- Whisper IME's readme says it provides an input method editor (keyboard) and can also be selected as system-wide voice input.
- Whisper IME's readme says voice recognition works entirely offline once the Whisper models have been downloaded.
- Microsoft says Microsoft 365 Dictate does not store your audio or transcribed text, but your speech is sent to Microsoft to produce the text.
- Microsoft says that when you use Dictate in Word, your speech utterances are sent to Microsoft and used only to provide text results.
- Microsoft's English command table for Dictate in Word lists 'delete or scratch or erase', 'delete that or scratch that' and 'undo' as commands.
- Microsoft says Dictation in Microsoft 365 uses speech-to-text with a microphone and a reliable internet connection.
- Microsoft says the dictation feature in Word is only available to Microsoft 365 subscribers.