IraVoice guides
Use your voice with fewer hidden assumptions.
Learn the built-in option first, understand where processing happens, then choose a dedicated tool only when it solves a real workflow constraint.
Three free private tools
Measure speed, score transcripts, and trace privacy.
Compare voice WPM, word accuracy, and your own typing baseline with one fixed passage; calculate WER, CER, and edit alignment; then audit where audio and text go across six separate boundaries. The tools do not request microphone permission or submit your working text.
Practical guides
Find the workflow you are trying to understand.
All 88 guides are grouped by the surface where speech becomes text. Every link remains visible, source-reviewed, and ready to browse without a search box.
10 source-reviewed guides
Mac dictation essentials
Start with the operating-system path: setup, shortcuts, commands, languages, troubleshooting, browser behavior, and processing boundaries.
Speech to text on Mac
Speech to text on a Mac starts with Dictation.
For live speech to text on a Mac, turn on Dictation in System Settings → Keyboard, choose a shortcut, place the cursor in a supported text field, and trigger Dictation. Before sensitive use, read the processing note shown under the Dictation setting for your language and configuration. Dictation is not the same as transcribing a saved audio file.
Read the guide →Speech-to-text assistive technology on Mac
Speech to text is one Mac input option—not a complete accessibility plan.
Start by naming the barrier and the exact task: entering prose, controlling the interface, reviewing written text, or avoiding a physical keyboard. Test Apple Dictation for focused text entry. Use Voice Control when spoken navigation and editing are also required. Add Read & Speak for read-back, or evaluate Full Keyboard Access, Sticky Keys, Slow Keys, and the Accessibility Keyboard when speech is not the right input. Run the real workflow in the real app before relying on it, and follow the relevant school, workplace, or service process for any formal accommodation decision.
Read the guide →Free dictation for Mac
The best free or no-subscription Mac dictation path is not always the built-in one.
For ordinary voice typing, turn on Apple Dictation under System Settings → Keyboard before downloading anything. On a macOS 26 Apple-silicon Mac, Megaphone is the strongest current permanent-free finished app for people who value MIT-licensed source, Apple SpeechAnalyzer transcription, on-device cleanup, a personal Dictionary, app context, and voice macros. Test Google AI Edge Eloquent when local audio or video file transcription and Voice Edit matter. Use Voice Control for spoken navigation, Google Docs Voice typing for document-specific commands, or whisper.cpp when you want an engine rather than a finished app. IraVoice has a 3-day trial but is not permanently free.
Read the guide →One-time purchase dictation apps for Mac
One-time purchase dictation apps for Mac, compared honestly.
Current one-time options include Voice Type at US $19.99, Whryte at US $24.99, VoiceInk from US $25, TypeVox at US $39, Sotto at US $49, Susurr at €19.99, and IraVoice at US $50. Apple Dictation is included with macOS, and Google AI Edge Eloquent is a current Mac no-cap, no-subscription baseline with an unpublished desktop price for English local dictation, Voice Edit, and file transcription. These are not interchangeable: compare the active speech model, optional cloud paths, supported Macs, device allowance, trial, updates, and whether you need live dictation or saved-file transcription.
Read the guide →Mac Dictation shortcut
The Mac Dictation shortcut is the one shown on your Mac.
Open System Settings → Keyboard, scroll to Dictation, and read the Shortcut value. Use that combination to start or stop Dictation. If your keyboard has a Microphone key, press and release it to start; holding it can activate Siri instead.
Read the guide →Mac Dictation troubleshooting
Mac Dictation not working? Diagnose the failure before replacing it.
Test Dictation in a new TextEdit document first. If it fails there, verify Keyboard → Dictation, the active shortcut, Voice Control, language, microphone source and input level, then check whether your configuration needs a network connection. If TextEdit works, isolate the target app or editing field.
Read the guide →Mac Dictation commands
Mac Dictation commands you can actually use while writing.
While standard Dictation is listening, say the name of the punctuation mark or formatting action: “comma,” “question mark,” “new line,” “new paragraph,” “no space on,” or “all caps.” Availability can vary by language, and Voice Control uses a different command system.
Read the guide →Mac Dictation languages
Change the Dictation language—and check which features travel with it.
Open System Settings → Keyboard → Dictation, choose Edit next to Languages, then select the language and region you want. With more than one language enabled, switch from the language control beside the Dictation cursor or press the Globe key and choose a language. Check Apple’s current macOS Feature Availability page separately for base Dictation, on-device and modeless Dictation, auto-punctuation, and dictated emoji.
Read the guide →Browser voice typing on Mac
A web text box and a website microphone are different voice paths.
For an ordinary supported web text field, click the field and start macOS Dictation with the shortcut shown under System Settings → Keyboard. That system path does not require granting the website microphone access. If the site has its own microphone button, Chrome or Safari and the individual website need microphone permission, and the site’s speech-processing terms apply.
Read the guide →Dictation privacy
Local versus cloud dictation, without the privacy theater.
Local dictation processes speech on your device. Cloud dictation sends audio to a remote service. Hybrid products split stages between the two. The useful question is not the label—it is what data leaves, for which stage, for how long, and under which setting.
Read the guide →34 source-reviewed guides
Writing and productivity
Choose the right input path for documents, notes, tasks, spreadsheets, and long-form writing without confusing dictation with recording or transcription.
Dictation software for writers on Mac
The best dictation setup for a writer starts with the manuscript—not the microphone.
Start with Apple Dictation in the writing app where the manuscript should live. Use Google AI Edge Eloquent when a free, fully local cross-app Mac workflow and file transcription fit better; Word Dictate or Google Docs Voice typing when their native commands matter; Scrivener or Ulysses when manuscript structure matters more than a separate speech engine; WriterDuet’s browser Dictate for its line-type commands; or Final Draft 13 on Mac for its English screenplay-aware Speech Commands. Dragon Professional remains a Windows product. Test one real scene, article section, or script page, then choose by correction burden, privacy, editor fit, revision safety, and export—not an unsupported headline accuracy number.
Read the guide →Google Docs voice typing
Google Docs voice typing works on Mac. The browser owns the speech path.
Open a document in the latest Chrome, Edge or Safari, choose Tools → Voice typing, allow microphone access, then click the microphone. Google says the browser controls the speech-to-text service and sends the resulting text to Google Docs.
Read the guide →Google Slides voice typing
Google Slides voice typing writes speaker notes—not slide text.
Open a Google Slides presentation in a current Chrome, Edge, or Safari browser, choose Tools → Dictate speaker notes, allow microphone access, and speak into the notes pane. Google’s native voice-typing control is for speaker notes; use an editable slide field with macOS Dictation or another text-input tool when the words must appear on the slide itself.
Read the guide →Google Sheets voice typing
Google Sheets does not have the Docs voice-typing menu.
Open the spreadsheet, click an empty cell or double-click a non-empty cell, and put the insertion point inside the editable value. Trigger the Dictation shortcut configured under macOS Keyboard settings, speak one cell’s content, stop Dictation, review the value or formula, then press Enter to commit it. Google does not document a Docs-style Tools → Voice typing command for Sheets.
Read the guide →Google Forms voice typing
Google Forms has text fields, not a Docs-style voice-typing button.
Open the form, click inside a Short answer or Paragraph response field, and confirm that the insertion point is visible. Start the Dictation shortcut shown under System Settings → Keyboard → Dictation, speak, stop, and correct the visible text. Resolve every required-field or response-validation error, then click Next or Submit yourself. Google does not document a Forms equivalent of Docs’ Tools → Voice typing command.
Read the guide →Excel voice typing on Mac
Excel voice typing has two Mac paths. Check which one you have.
Open the workbook and check the current Excel ribbon for Dictate while signed in. Microsoft’s Microsoft 365 overview names desktop Excel, a microphone, and reliable internet, but does not provide an Excel-specific procedure on that page. If Dictate is unavailable, press F2 to edit the selected cell or place the insertion point in the Formula bar, then trigger the Dictation shortcut configured under macOS Keyboard settings. Speak one value, stop, inspect the value or formula, and press Return only when it is correct.
Read the guide →PowerPoint dictation on Mac
PowerPoint Dictate writes slides and notes. It does not record narration.
Open PowerPoint, sign in to Microsoft 365, choose Home → Dictate, wait for listening to start, then move the insertion point to a slide placeholder or the notes pane and speak. Use Record when you need saved narration, Speaker Coach for rehearsal feedback, or subtitles when the audience needs live on-screen text.
Read the guide →Microsoft Word dictation
Microsoft Word has Dictate on Mac. Start with the tool Word already owns.
Open a Word document, sign in to Microsoft 365, choose Home → Dictate, wait for the control to begin listening, and speak. Microsoft says Word Dictate requires a microphone, an active Microsoft 365 subscription, and a reliable internet connection.
Read the guide →OneNote dictation
OneNote Dictate works on Mac. OneNote Transcribe is a different feature.
In OneNote for Microsoft 365 on Mac, sign in, click an editable page, then choose Home → Dictate. Microsoft requires a microphone and reliable internet connection for this connected speech-to-text service. Speak after the Dictate control starts listening; use its gear menu to choose the language, microphone, automatic punctuation, and profanity filter.
Read the guide →Notion voice typing
Notion has desktop voice input. Know which field receives your words.
Use the voice control exposed in Notion Agent when you want to speak an AI prompt. To enter ordinary text directly into a Notion page, place the insertion point in an editable block and use Apple Dictation or another Mac-wide dictation tool.
Read the guide →Apple Notes dictation
Dictating text and transcribing audio in Apple Notes are different jobs.
For ordinary text, click in a Notes text field and trigger the Dictation shortcut shown in System Settings → Keyboard. To preserve audio and create a transcript, choose Attachment → Record Audio inside an iCloud note; Apple currently requires an M1-or-later Mac and a supported device language for Notes audio transcription.
Read the guide →Notability dictation on Mac
In Notability, voice typing and audio transcripts create different things.
To voice type editable text in Notability on a Mac, select the Text tool, click in the note until a real insertion point appears, then use the Dictation shortcut shown under System Settings → Keyboard. To keep audio and generate a time-stamped transcript, start Notability recording with its microphone control or Command-R and use the transcript feature available to the current plan. These are separate workflows.
Read the guide →Goodnotes dictation on Mac
Goodnotes voice typing and audio transcription are separate workflows.
To voice type editable text in Goodnotes on a Mac, open a notebook text box or Text Document, place the insertion point where the words should appear, then use the Dictation shortcut shown under System Settings → Keyboard. To preserve audio and create a Goodnotes transcript, use Record and Summarize in a supported notebook and check the current transcription mode and plan shown in the app. These paths do not create the same artifact.
Read the guide →Apple Pages speech to text
Use speech to text in Apple Pages with Mac Dictation.
To use speech to text in Apple Pages on a Mac, open a document, place the insertion point where the words should appear or select text to replace, then choose Edit → Start Dictation. You can also use the Dictation shortcut shown under System Settings → Keyboard. Speak when the microphone appears, stop Dictation, and review the editable text.
Read the guide →LibreOffice dictation on Mac
LibreOffice Writer has editable text—not a native Dictate button.
Open a writable LibreOffice Writer document, click in the document body until the insertion point appears, then start the shortcut shown under System Settings → Keyboard → Dictation. Speak a short passage, pause for the text to appear, stop Dictation, and review it before applying styles or saving. LibreOffice does not document a built-in Writer Dictate button; macOS owns this speech-to-text path.
Read the guide →Adobe Acrobat dictation on Mac
Acrobat dictation works only after you find a real text surface.
Open the PDF in Acrobat or Acrobat Reader, identify whether you are filling an interactive field, adding a Fill & Sign text overlay, writing a comment, or editing document text, and place the insertion point in that exact surface. Start the shortcut shown under System Settings → Keyboard → Dictation, speak, stop, and review the visible text before saving, posting, signing, or submitting anything.
Read the guide →Todoist voice input on Mac
Todoist Ramble creates tasks. Dictation writes into the field you already chose.
To use Todoist Ramble on Mac, choose the waveform in Quick Add or the sidebar, or press Shift + Q while Todoist is active. For new desktop installs, Todoist documents Option + Shift + R as the global Quick Ramble shortcut; older installs may still use Shift + Option + Space, so check Settings → Desktop. Speak each task clearly, review the live preview, make spoken corrections if needed, then finish with “that’s all” or the Add tasks control. For prose inside an existing description or comment, use focused-field Dictation instead.
Read the guide →Grammarly voice typing on Mac
Grammarly’s voice typing is on iOS. On Mac, dictate first and review second.
Install and enable Grammarly for Mac, open a supported app or website, and place the insertion point in the intended text field. Start the Dictation shortcut shown under System Settings → Keyboard or use another Mac input tool, speak, then stop. When the text appears, review Grammarly’s underlines or open the Grammarly tab and accept or dismiss each relevant suggestion. Grammarly’s native speech-to-text control is currently documented for Grammarly Keyboard on iOS, not Grammarly for Mac.
Read the guide →Scrivener dictation on Mac
Scrivener uses the Mac’s dictation path. Your Binder still decides where the words belong.
Open the intended Scrivener project, select the correct document in the Binder, and click in its editor until the insertion point is visible. Confirm Dictation is enabled under System Settings → Keyboard, then use the shortcut shown there and speak. Stop Dictation, inspect punctuation and paragraph breaks, and edit the draft before moving to another document. For a large spoken revision, create a Scrivener snapshot first; use File → Compile only when you want to assemble selected project documents into a separate output file.
Read the guide →WriterDuet dictation on Mac
WriterDuet’s screenplay-aware Dictate mode is browser-only.
For WriterDuet’s native Dictate mode, sign in at writerduet.com/script from a computer web browser, open the correct project and document, then choose Edit → Dictate or press Command-Shift-D. Speak script text and use “Press enter,” “Press enter character,” or another documented line-type command. Stop Dictate and review every line before sharing or exporting. In the Mac desktop app, WriterDuet’s native Dictate mode is unavailable; use macOS Dictation or another focused-field input tool only after testing that the editor accepts a short phrase.
Read the guide →Final Draft dictation on Mac
Final Draft’s Mac dictation can structure a screenplay—not just insert prose.
On a Mac, enable Dictation in System Settings → Keyboard, select and test the intended microphone, then open Final Draft 13 and enable Speech Commands under Final Draft 13 → Preferences → General. Place the cursor in the correct script, start from Edit → Start Dictation, and speak a short test. Use commands such as “return,” “tab,” or “new paragraph… scene heading” only after confirming the English command layer is active. Stop dictation, review both the words and screenplay elements, save the script, and use Speech Control separately when you want read-aloud proofreading.
Read the guide →Sudowrite dictation on Mac
Sudowrite has Smart Dictation—but the native microphone is mobile-only.
For Sudowrite’s native Smart Dictation, use the iOS or Android companion app: start New Dictation from the plus menu or tap the sparkly microphone in an open document, speak, stop, review the formatted result, and deliberately copy it or add it to a project. On a Mac, open the intended project and document at sudowrite.com, click the exact insertion point, trigger the Dictation shortcut shown under System Settings → Keyboard, speak, stop, and review before using Write, Rewrite, Story Bible, collaboration, or export.
Read the guide →Substack voice to text on Mac
Substack has several voice workflows. Only one puts editable words at your Mac cursor.
For editable Substack text on Mac, open the correct publication Dashboard, choose Create → Article, click in the post body, and start the Dictation shortcut shown under System Settings → Keyboard. Speak, stop, and review the draft before selecting Continue. Then confirm the audience, comments, free preview, tags, email and app delivery, and schedule before publishing. Use Substack’s Voiceover, audio embed, podcast transcript, or reader text-to-speech only when that separate audio artifact is actually the goal.
Read the guide →WordPress voice to text on Mac
WordPress voice typing starts in the editor. A speech plugin solves a different problem.
Open the intended WordPress site and post, click inside a Paragraph or other editable text block, and start the Dictation shortcut shown under System Settings → Keyboard. Speak, stop, and review the text before applying block structure. Save the draft or preview it, then verify the title, permalink, categories, tags, featured image, visibility, schedule, and final Publish action. A speech-to-text plugin is not required for basic Mac author dictation; it is additional site software for a separate editor or visitor-facing requirement.
Read the guide →TickTick voice input on Mac
TickTick can turn speech into tasks. On Mac, choose the Shortcut before the microphone.
For native TickTick voice-to-text on Mac, use TickTick 4.0.21 or later and build an Apple Shortcut with Dictate Text followed by TickTick’s Add Content to Task action. Map the dictated result to Title, choose the intended List, run the Shortcut, then review the task and any Smart Recognition date or time before saving. AI Voice Add is currently documented as mobile only; direct Siri task creation must first be enabled under Settings → Import & Integration → Siri.
Read the guide →Evernote voice to text on Mac
Evernote now dictates directly into notes. Recording and AI transcription remain different paths.
For live Evernote voice-to-text on Mac, update to Evernote 11.21.5 or later, open the intended note, place the insertion point where the text belongs, then use Evernote’s direct dictation control and review the result. Use Voice note or Meeting recorder when audio should remain attached to the note, and use AI Transcribe when an existing audio, video, image, or supported URL should become editable text. Evernote’s release note does not currently document the direct-dictation shortcut, model, local-processing status, retention rule, or language list, so do not borrow those claims from its separate AI Transcribe documentation.
Read the guide →Google Keep voice typing on Mac
Google Keep voice typing on Mac is a focused text-field workflow.
Open keep.google.com, click Take a note, place the insertion point in the title or note body, and start your configured macOS Dictation shortcut. Speak, stop Dictation, review the text, then click Done separately. If you want a saved audio note with playback, use Google Keep’s documented Android Recording workflow instead.
Read the guide →Obsidian voice-to-text on Mac
In Obsidian, voice-to-text and audio recording are different jobs.
For live Obsidian voice-to-text on a Mac, open a note in Live Preview or Source mode, place the cursor in the editor, and start the Dictation shortcut configured under System Settings → Keyboard. Use Obsidian’s core Audio recorder when you want to save and embed the recording itself. A community transcription plugin is a separate trust and processing decision.
Read the guide →Logseq voice to text on Mac
Dictate the thought into Logseq. Build the graph deliberately.
Open the intended Logseq graph, go to today’s Journal or the correct page, click an editable block until the insertion point appears, and trigger the Dictation shortcut shown under System Settings → Keyboard. Speak the prose, stop, and review it. Use Logseq’s keyboard shortcuts and slash commands for new blocks, page references, properties, task markers, priorities, deadlines, and other graph structure instead of assuming spoken formatting will create valid Logseq syntax.
Read the guide →Ulysses dictation on Mac
Ulysses already uses Apple Dictation inside the editor.
Open the intended Ulysses sheet, click in the editor, and trigger the Dictation shortcut shown under System Settings → Keyboard. Ulysses documents Apple’s system Dictation path; stop and review the prose, then apply Markdown XL structure deliberately. Choose an iCloud-synced or On My Mac location before sensitive work and export only the sheets you intend to share.
Read the guide →Joplin voice typing on Mac
Joplin documents built-in voice typing for Android—not its Mac app.
Open the intended Joplin note on your Mac, click inside the Markdown or Rich Text editor until the insertion point appears, and trigger the Dictation shortcut shown under System Settings → Keyboard. Speak, stop, and review the note before syncing. Joplin’s documented built-in offline voice typing uses Whisper.cpp or Vosk on Android; that Android feature should not be presented as a Mac desktop control.
Read the guide →Drafts dictation on Mac
Drafts has a real Mac dictation interface—and a separate file transcription tool.
For live speech on Mac, choose File → New with Dictation to create a draft or Editor → Dictate to insert into the current draft; Drafts also supports the normal macOS Dictation hot key in its editor. Speak, review the preview, and choose Create or Insert before assuming the text is saved. For an existing recording, use Editor → Transcribe instead and select the correct language or locale before choosing the file.
Read the guide →Day One dictation on Mac
On Mac, dictate journal text into Day One—but do not expect the iOS audio recorder.
Create or open the intended Day One entry on Mac, click in the editable body, and trigger the Dictation shortcut shown under System Settings → Keyboard. Speak, stop, and review the journal text before leaving the entry or syncing it. Day One’s built-in audio recording is not currently available on macOS, and its automatic audio-to-text feature is documented as iOS-only; adding an existing audio file to a Mac entry does not make it transcribable in Day One.
Read the guide →Anki voice input on Mac
Anki voice input starts by choosing text, audio, or review.
To dictate editable flashcard text in Anki Desktop on a Mac, open Add or edit an existing note, click the intended Front, Back, Extra, cloze, or custom field until an insertion point appears, then use the Dictation shortcut shown under System Settings → Keyboard. Stop Dictation and review the text before adding or updating the note. Anki’s microphone button records and attaches audio; it is not the live speech-to-text control.
Read the guide →16 source-reviewed guides
AI and developer tools
Separate native voice features, terminal safety, editor input, and agent submission so spoken text stays reviewable before it becomes an action.
Terminal and iTerm voice typing on Mac
Voice typing in Terminal should create a draft—not execute a guess.
For a harmless prose or agent prompt, click the active Terminal or iTerm2 command line, start the Dictation shortcut shown under System Settings → Keyboard, speak, stop Dictation, and inspect the complete unsent text before pressing Return. Do not voice-type destructive commands, secrets, or exact shell syntax blindly. If direct insertion is uncertain, dictate into TextEdit or another plain-text editor, verify every character, then paste deliberately; iTerm2 users should also review Paste History settings.
Read the guide →Warp terminal voice input
Warp already has native voice input. Start there.
In Warp, open Settings → Agents → Warp Agent → Voice, enable Voice, and choose an activation hotkey. On macOS, Warp currently documents holding Fn to record and releasing it to stop and transcribe; Agent Mode also has a microphone button. Warp says Wispr Flow powers transcription, processes voice data in real time, and does not retain it as a recording after transcription. Review the resulting prompt or command before submission or execution.
Read the guide →Claude Desktop voice input on Mac
Claude Desktop has Caps Lock dictation. Voice Mode is a different path.
In Claude Desktop on Mac, open Settings → General, enable the optional voice shortcut, then press Caps Lock once to start dictating and again to stop. Review the real-time transcript and click the arrow to send it. Anthropic currently requires macOS 14 or later, Claude Desktop running in the background, and Speech Recognition permission for this Quick Entry voice path.
Read the guide →Gemini Mac voice mode
Gemini for Mac is live. Its voice-edit workflow is still marked coming soon.
Install Gemini for Mac from Google, sign in, and use Option + Space for the documented mini chat or Option + Shift + Space for full chat. Google’s current product page still labels its screen-aware “Think out loud” voice-edit experience Coming soon. Until the native voice control appears in your account, place the cursor in a supported Gemini text field and use the current macOS Dictation shortcut, then review the text before sending.
Read the guide →Claude Code
Claude Code already has voice dictation. Here is when that is enough.
In Claude Code v2.1.69 or later, run /voice and hold Space to dictate. Tap mode is available in newer builds. Anthropic says audio is streamed to its servers for transcription and the feature requires a Claude.ai account.
Read the guide →GitHub Copilot CLI voice input
Copilot CLI has local voice input. Start there.
Start an interactive Copilot CLI session, run /voice, and download the voice runtime plus an English or Spanish model. Hold Space for a short prompt, or press Control-X then V for toggle recording. GitHub says transcription runs locally and the audio is not sent over the network.
Read the guide →GitHub voice typing on Mac
GitHub’s website gives you editors—not a microphone.
On github.com, place the insertion point in an editable issue title or body, pull-request description, discussion, comment, or review field. Start the Dictation shortcut configured under macOS Keyboard settings, speak, stop Dictation, and review the visible text and Markdown preview. Submit with GitHub’s explicit button only after checking repository, mentions, references, links, code identifiers, and review status.
Read the guide →NotebookLM and Gemini Notebook voice input on Mac
NotebookLM is now Gemini Notebook. Voice input is still a text-field job.
Open NotebookLM or Gemini Notebook in your browser, choose the sources the answer should use, and place the insertion point in the Chat prompt box. Start the Dictation shortcut configured under macOS Keyboard settings, speak one source-grounded question, stop Dictation, and review the visible prompt before sending. Source selection, citations, response settings, Save to note, and Audio Overview remain separate Gemini Notebook controls.
Read the guide →JetBrains and IntelliJ voice input on Mac
JetBrains Air has local voice input. Classic JetBrains IDEs are a different path.
In JetBrains Air, open Air Settings → Voice Input, download a local Whisper model, return to Chat, and select the microphone in the task input. In IntelliJ IDEA, PyCharm, WebStorm, or another classic JetBrains IDE, focus the intended editor, comment, documentation, commit-message, or AI Assistant prompt field and use the macOS Dictation shortcut when that field accepts it. Review code identifiers, punctuation, context, agent, permissions, and every Send, Commit, or run action separately.
Read the guide →Gemini CLI voice mode
Gemini CLI voice can be cloud or local. Choose before speaking.
Update Gemini CLI to a release with Voice Mode, open /settings, and enable the experimental Voice Mode setting. Run /voice, then hold Space in the default push-to-talk mode. Before sensitive use, check Voice Transcription Backend: Gemini Live sends recordings to Google Cloud, while Whisper performs transcription locally.
Read the guide →ChatGPT dictation
ChatGPT Dictation creates editable text. Voice talks back.
Select the microphone icon in ChatGPT when you want to dictate one prompt. ChatGPT sends the recorded audio to OpenAI’s models for transcription, then returns editable text that you can review before sending. Choose the separate Voice control for a live spoken conversation, or Record in an eligible ChatGPT macOS app when the job is meeting or voice-note transcription.
Read the guide →Codex
Codex has live Voice. IraVoice is a different kind of input.
Open the ChatGPT desktop app, choose Codex, select the Voice control, allow microphone access, and speak. For a local text-first workflow, use IraVoice to produce editable Markdown for focused delivery or an explicit, guarded Codex handoff.
Read the guide →VS Code
VS Code has private native dictation. Start there.
Install the official VS Code Speech extension, press Option-Command-V to dictate in the editor, or press Command-I to start voice chat. Microsoft says recordings are processed locally and are not sent to an online service.
Read the guide →Cursor
Cursor has voice input. Know what it does before adding another tool.
In current Cursor builds, press and hold Control-M to record, then use the waveform, timer, cancel, and confirm controls. Cursor describes batch speech-to-text but does not document the voice-specific processing path on that changelog page.
Read the guide →Windsurf and Devin Desktop voice input
Windsurf voice works in Cascade. That is not editor-wide dictation.
Use the Voice input control in Cascade when the current Cascade conversation is the destination. The official documentation says it transcribes speech to text, but the reviewed page does not name a voice shortcut, speech provider, audio-retention period, or local-processing guarantee. For another editable field on a Mac, focus that field and use macOS Dictation or a separate local dictation tool, then review the text before an agent or terminal receives it.
Read the guide →Zed editor voice input
Zed voice chat is shipped. Assistant dictation is still planned.
Do not treat the microphone in a Zed collaboration channel as dictation: Zed documents it as voice chat with other people. The current roadmap lists Dictation for Assistant as Planned and a newer Voice Input request in triage, so verify the live release before following any future setup. For editable text now, focus the intended Zed editor or agent field and use macOS Dictation or another reviewed system-wide Mac input path.
Read the guide →17 source-reviewed guides
Messages, email, and meetings
Put words in the correct composer while keeping recipients, send actions, recordings, captions, and saved transcripts as separate decisions.
Email dictation on Mac
Dictate the email where you write it—but review before sending.
In Apple Mail or Gmail, place the insertion point in the message body and start the Dictation shortcut configured under System Settings → Keyboard → Dictation. In Outlook for Mac, Microsoft also documents Message → Dictate and the Option-F1 shortcut for Microsoft 365 subscribers. Review recipients, names, numbers, links, tone, and confidential details before you choose Send.
Read the guide →Gmail voice typing on Mac
Gmail has no Docs-style Voice typing menu. Use the Mac input path.
Open Gmail on your Mac, choose Compose or Reply, click in the message body, then start the Dictation shortcut shown under System Settings → Keyboard. Speak, stop Dictation, and verify recipients, subject, names, numbers, links, attachments, and tone before clicking Send. Google Docs’ Tools → Voice typing control belongs to Docs and Slides; it is not a Gmail compose command.
Read the guide →Apple Mail dictation on Mac
Mail can receive Mac Dictation. Sending is still a separate action.
In Apple Mail on Mac, choose New Message or Reply, click in the message body, then start the Dictation shortcut shown under System Settings → Keyboard. Speak, stop Dictation, and verify recipients, subject, names, dates, amounts, links, attachments, and tone. Save the message as a draft when you need more review; click Send only as a separate deliberate action.
Read the guide →Outlook dictation on Mac
Outlook has its own Dictate button. Know what it sends.
In Outlook for Mac, start a New Email or reply, click the message body, then choose Message → Dictate or press Option-F1. Microsoft says the feature requires a Microsoft 365 subscription, sign-in, microphone, and reliable internet connection. Speech utterances are sent to Microsoft for transcription; Microsoft says it does not store the audio data or transcribed text for this service. Stop Dictate, verify recipients, names, numbers, links, attachments, and tone, then send the message yourself.
Read the guide →Slack voice typing on Mac
Slack has three voice paths. Only one writes an ordinary message.
For editable channel, thread, or DM text on a Mac, click the Slack message field and start the Dictation shortcut configured under System Settings → Keyboard. Slack separately offers microphone dictation inside Slackbot on eligible plans, while its microphone button in a conversation records an audio clip rather than composing an ordinary text message.
Read the guide →Discord voice typing on Mac
Discord voice typing starts in the message box—not the voice channel.
For editable Discord text on a Mac, open the intended text channel, DM, thread, or voice-channel text chat; click the message box; and start the Dictation shortcut configured under System Settings → Keyboard. Speak, stop, review the visible draft, and explicitly press Enter or choose Send. Discord’s official Voice Messages are saved audio that can currently be sent only from the mobile app, while /tts reads typed text aloud.
Read the guide →WhatsApp voice typing on Mac
WhatsApp voice typing should create the message you mean to send.
For ordinary editable WhatsApp text on a Mac, open the intended chat, click the message field, and start the Dictation shortcut configured under System Settings → Keyboard. Speak, stop, review the visible draft, and explicitly send it. WhatsApp’s microphone control records a voice message instead. Voice-message transcripts turn received audio into text on supported devices, but WhatsApp’s current Mac Help Center says the feature is available only on certain devices and may not be available to you yet.
Read the guide →Facebook Messenger voice typing on Mac
Messenger voice typing starts in the text box—not the audio recorder.
For a new editable Facebook Messenger message on a Mac, open Messenger.com in a current supported browser, choose the intended conversation, click the text box, and use the Dictation shortcut configured under System Settings → Keyboard. Stop, review the visible draft and recipient, then explicitly send it. The microphone beside the text box records a voice message instead. Meta’s current first-party Messenger pages reviewed here document recording and sending audio, but do not provide a general Messenger.com procedure for transcribing an existing voice message.
Read the guide →Instagram voice typing on Mac
Instagram voice typing starts with the text you want to create.
To voice type a new Instagram DM on a Mac, open Instagram.com, choose the intended conversation, click the message field, and start the Dictation shortcut configured under System Settings → Keyboard. Speak, stop, review the visible draft and recipient, then explicitly send it. Instagram voice messages preserve audio instead. Some current mobile accounts show View transcription below an individual voice message, while Meta’s public Help pages reviewed here do not provide a stable computer procedure. Reel captions, voiceovers, and text to speech are separate mobile creation paths.
Read the guide →TikTok voice to text on Mac
TikTok voice to text depends on which text you want.
To speak new editable TikTok text on a Mac, focus a text field that the current TikTok web experience actually exposes to your account, start the Dictation shortcut configured under System Settings → Keyboard, speak, stop, review the visible draft, and only then post or send it. TikTok Studio documents a browser Upload workflow with a Caption control, but tool availability varies between web and app. TikTok’s auto-generated and creator captions instead turn speech already inside a video into text; text to speech runs in the opposite direction; voice messages preserve audio; and a public-video transcript generator is a separate third-party service.
Read the guide →iMessage and Messages dictation on Mac
Dictate the text. Keep the recipient and Send action visible.
To dictate an editable iMessage or text message on a Mac, open the intended conversation, click the message field, start the Dictation shortcut configured under System Settings → Keyboard, speak, stop Dictation, review the words and recipient, then press Return to send. Apple’s Record Audio control instead sends retained audio with an automatic transcript, and Siri’s “Message…” command is a separate hands-free compose-and-send workflow.
Read the guide →Signal voice typing on Mac
Signal voice typing starts in the composer—not the recorder.
For a new editable Signal message on a Mac, open the intended chat, focus the composer with Command-Shift-T or click the message field, and start the Dictation shortcut configured under System Settings → Keyboard. Speak, stop, review the visible draft and recipient, then explicitly send it. Command-Shift-Y begins Signal’s separate voice-message recorder. Signal’s current first-party pages reviewed here do not document a native voice-message transcription control; forwarding or exporting audio to a bot or converter makes that third party another content recipient.
Read the guide →Telegram voice typing on Mac
Telegram voice typing and voice-to-text solve opposite problems.
For an ordinary editable Telegram message on a Mac, open the intended chat, click the message field, and use the Dictation shortcut configured under System Settings → Keyboard. Speak, stop, review the visible draft, and explicitly send it. For an existing eligible voice or video message, Telegram’s separate Voice-to-Text feature uses the →A control. Current Telegram documentation exposes a limited non-Premium trial and unlimited Premium conversion, and Telegram says selected audio may be shared with Google to produce the transcript.
Read the guide →Microsoft Teams voice typing
Teams has five voice artifacts. Pick the one you actually need.
For ordinary editable Teams chat text on a Mac, click the compose box and use the Dictation shortcut configured under macOS Keyboard settings. Microsoft’s current Teams chat instructions document composing text in that box but do not document a built-in dictation control for an ordinary chat. A Teams audio clip sends up to one minute of recorded audio; live captions are not saved; meeting transcripts are stored and downloadable; and Microsoft 365 Copilot Chat has its own Start dictation button.
Read the guide →Zoom voice typing on Mac
Zoom voice typing can mean chat text, captions, or a saved transcript.
For an ordinary editable Zoom Chat or in-meeting chat message on a Mac, click the compose box and use the Dictation shortcut configured under System Settings → Keyboard. Stop, review the visible draft and destination, then explicitly send it. Zoom’s automated captions display meeting speech in real time; Meeting transcripts are designed to retain and save speech-to-text data; and cloud-recording audio transcription produces a separate VTT file after a qualifying recorded meeting.
Read the guide →Google Meet voice to text on Mac
Google Meet voice to text starts by choosing the artifact.
For an editable Google Meet chat message on a Mac, focus the chat composer and use the Dictation shortcut shown under System Settings → Keyboard, then review and send the draft yourself. Use Meet live captions to read current speech, Transcripts to retain spoken words in the organizer’s Drive, or Take notes with Gemini to create a shared summary. These features are not interchangeable.
Read the guide →Loom voice to text on Mac
Loom voice to text begins with a recording—not a blank text field.
Record with Loom when the video and its transcript are the intended artifacts. Loom says transcripts are available on all plans, support more than 50 languages, and can be copied or searched; caption downloads and transcript editing depend on plan and role. For ordinary editable text around a Loom, focus the field and use macOS Dictation, then review before publishing or sharing.
Read the guide →9 source-reviewed guides
Professional work, CRM, and support
Dictate into legal, project, customer, and support systems with the data path, field, permissions, visibility, and workflow state still under review.
Dictation software for lawyers on Mac
For legal dictation, “local” is a data-path fact—not a privilege verdict.
On a Mac, test Apple Dictation first in the exact Word, Outlook, PDF, browser, or practice-management field used for legal work, then read the processing disclosure shown under System Settings → Keyboard → Dictation. Use Microsoft 365 Dictate only when its connected speech service and account controls fit the organization’s review. Nuance’s current Dragon Legal v16 page identifies Windows 10 and 11—not macOS—while Dragon Legal Anywhere is cloud-hosted. Treat IraVoice or any other local Mac app as one candidate data path, never as proof of privilege, confidentiality, compliance, or professional approval.
Read the guide →Jira voice to text on Mac
Jira has rich text fields. Voice input still comes from somewhere else.
In Jira Cloud on a Mac, press C or select Create, focus the intended Title, Description, or comment field, and start the Dictation shortcut configured under macOS Keyboard settings. Speak, stop Dictation, and review the visible field text plus the selected space, work type, required fields, restrictions, and inherited values. Select Create or save only after that review.
Read the guide →Asana voice to text on Mac
Asana’s native voice task is on iOS. Mac text fields are a different workflow.
On a Mac, open the intended Asana task or create-task form, focus the exact name, description, or comment field, and start the Dictation shortcut configured in macOS. Speak, stop, and review the text plus project, assignee, due date, collaborators, custom fields, and access before selecting Create, saving the task, or posting the comment. Use Asana’s native voice-task flow on iOS when you want its automatic title, description, and attached-audio workflow.
Read the guide →ClickUp voice to text on Mac
ClickUp has three voice workflows. Only one types wherever your cursor is.
For ClickUp voice to text across Mac apps, keep the Brain MAX desktop app open, focus the exact text box, hold fn or your configured Talk to Text shortcut, speak, then release to stop. ClickUp says it converts the speech to text, improves it with AI, and pastes the result into the current field. Inside ClickUp, you can instead use Talk to Text where its microphone appears. Use Voice Clips only when you intend to send and retain an audio comment.
Read the guide →Zoho CRM voice to text on Mac
Zoho CRM has four voice features. None of them is generic text input everywhere.
For an ordinary Zoho CRM note on Mac, open the intended module and record, focus the Notes text box, start the Dictation shortcut shown under System Settings → Keyboard, speak, stop, review the record plus note title and content, then click Save. Use Zia Voice Assistant only for its predefined navigation and action commands, activated with Control + Z + V or the bottom-bar microphone after enabling it. Ask Zia voice input belongs to its conversational Insights, Help, and Action interface; mobile voice-to-text notes and recorded-call transcription each have separate storage and commitment boundaries.
Read the guide →Confluence voice typing on Mac
Confluence has speech-to-text for Rovo creation. Ordinary editor text is a different path.
For ordinary Confluence voice typing on Mac, open the intended page, live doc, or comment, focus the exact editable field, then start the Dictation shortcut configured under System Settings → Keyboard. Speak, stop, and review the text plus its audience before publishing, updating, or commenting. Use Create with Rovo speech-to-text only when the goal is an AI-generated Confluence draft or canvas, and use Loom only when the recording and its transcript should remain a separate shared artifact.
Read the guide →Salesforce voice to text on Mac
Salesforce has several voice-to-text paths. A Mac record field is only one of them.
For ordinary Salesforce voice-to-text on Mac, open the intended Lightning Experience object and record, click Edit or the pencil beside an editable text field, focus the exact field, then start the Dictation shortcut configured under System Settings → Keyboard. Speak, stop, review the record, field, transcript, validation, and any save options, then click Save. Do not use that workflow as a substitute for Agentforce mobile voice actions, Mobile AI Transcription, a configured Consumer Goods flow, stored-file transcription in Flow Builder, or Salesforce Voice telephony.
Read the guide →HubSpot voice to text on Mac
HubSpot has mobile Breeze voice input. Mac CRM fields are a different path.
For a HubSpot CRM note on Mac, open the intended record, choose Note or the correct activity, focus its text field, then start the Dictation shortcut configured in macOS. Speak, stop, and review the transcript plus record associations, activity type, visibility, date, outcome, and other fields before selecting Save note or Log activity. Use HubSpot mobile Breeze voice input when the actual goal is speaking a prompt to Breeze Assistant.
Read the guide →Zendesk voice to text on Mac
In Zendesk, the dangerous question is not transcription. It is who can see the reply.
Open the intended Zendesk ticket, choose Public reply or Internal note before composing, focus the exact text field, and start the Dictation shortcut configured in macOS. Speak, stop, and review the transcript, channel, recipients, attachments, ticket fields, and visibility. Use Zendesk Draft mode where available, then submit only after the public-reply confirmation.
Read the guide →2 source-reviewed guides
Design and creative
Identify the editable text surface first, then keep live dictation distinct from media uploads, captions, and audio conversion.
Canva voice typing on Mac
Canva voice typing starts in an editable text box—not an audio converter.
To voice type editable text in Canva on a Mac, double-click the intended text box until the insertion point appears, then start the Dictation shortcut configured under System Settings → Keyboard. Speak, stop, and review the text before leaving the box. Use Canva’s Transcribe Audio or Speech to Text tools when the source is an existing audio or video file—not when you want live words at the cursor.
Read the guide →Figma voice input on Mac
Figma voice input depends on where the words should go.
Use Figma Make’s native voice-to-text control when speech should become a prompt that generates or revises a prototype. For editable Figma Design copy, enter text-edit mode by double-clicking the layer or pressing Enter, then use the Dictation shortcut configured under macOS Keyboard settings. Dictate comments inside the comment field. Figma audio chat and its closed captions are separate live-collaboration tools, not saved design text.
Read the guide →Editorial standard
Teach the native option before selling another app.
Every guide links to the first-party documentation behind its setup, privacy, and availability claims. Product details carry a visible review date.
IraVoice is recommended only where its actual boundaries—local Mac processing, reusable text, or Voice-to-Spec—solve a distinct problem.