- Spoken punctuation
- You name a mark, such as "comma", and a supported dictation tool inserts it.
- Automatic punctuation
- The engine adds punctuation from speech and language context without requiring spoken commands.
- Dictation engine
- The keyboard, operating system, or app that turns your speech into text and interprets commands.
01 / Say the mark
Master Voice Punctuation Cheat Sheet
Speak a supported punctuation command where the mark belongs, without stopping dictation. These are common English commands, not a universal command set.
Support varies by app and language, especially for quotation marks, parentheses, symbols, and paragraph breaks. A keyboard that recognizes "comma" does not necessarily support "open quote".
| What you want | What to say | Expected result |
|---|---|---|
| Period or full stop | "period" or "full stop" | . |
| Comma | "comma" | , |
| Question mark | "question mark" | ? |
| Exclamation point | "exclamation point" or "exclamation mark" | ! |
| New line | "new line" | Moves the following text to the next line |
| New paragraph | "new paragraph" | Starts a separate paragraph |
| Opening quotation mark | "open quote" or "quote" | " |
| Closing quotation mark | "close quote" or "end quote" | " |
| Colon | "colon" | : |
| Semicolon | "semicolon" | ; |
| Opening parenthesis | "open parenthesis" | ( |
| Closing parenthesis | "close parenthesis" | ) |
| Hyphen | "hyphen" | - |
| Apostrophe | "apostrophe" | ' |
Google Docs' documented native punctuation commands are more limited than this full table. If a symbol command is unsupported, insert it with the keyboard.
Parentheses and quotation marks need explicit commands
Say "open parenthesis" before the enclosed text and "close parenthesis" afterward. Saying only "parentheses" does not specify which mark to insert.
Quotation marks may appear as straight or typographic marks depending on the app. Opening and closing quote commands also vary, so test them before dictating a long quoted passage.
Example: dictate two punctuated paragraphs
Hello comma how are you question mark New paragraph I am doing well period
Hello, how are you?
I am doing well.
02 / Find the right switch
How to Enable Automatic Punctuation on Your Device
Turn on automatic punctuation in the keyboard, operating system, or dictation toolbar that is capturing your speech. It is an engine setting, not a universal microphone feature, and availability depends on the device, language, and software version.
| Device or app | Availability | Settings path |
|---|---|---|
| iPhone / iPad | Where supported | Settings > General > Keyboard > Auto-Punctuation |
| Android / Gboard | Selected configurations | Gboard Settings > Voice typing > Auto-punctuation |
| Windows 11 | Voice typing setting | Win + H > Settings gear > Auto-punctuation |
| Microsoft Word | Where available | Home > Dictate > Settings gear |
| Google Docs native voice typing | No native switch | Use spoken commands or a separate dictation engine |
Scroll the table horizontally to compare all columns.
iPhone and iPad: turn on Auto-Punctuation
Path: Settings > General > Keyboard > Auto-Punctuation
- Open Settings.
- Select General > Keyboard.
- Turn on Enable Dictation if it is off.
- Turn on Auto-Punctuation, where available.
- Open Messages, Notes, or another text field.
- Tap the keyboard's microphone and speak normally.
Automatic punctuation adds marks such as commas, periods, and question marks without requiring you to name them. You can still speak punctuation commands when you want explicit control.
Use the keyboard microphone for text. A messaging app's separate audio-recording button sends a voice message rather than typing your speech.
If the Auto-Punctuation toggle is missing
Check your device's supported dictation features, selected language, and iOS or iPadOS version. Apple does not offer every dictation feature in every language or on every device.
Android with Gboard: check Voice typing settings
Path on supported configurations: Gboard Settings > Voice typing > Auto-punctuation
- Tap a text field to display Gboard.
- Open the keyboard's Settings using its gear icon or toolbar menu.
- Select Voice typing.
- Turn on Auto-punctuation if the option appears.
- Return to your text field, tap Gboard's microphone, and dictate a test sentence.
Here is the catch: Gboard does not expose an Auto-punctuation toggle on every Android phone. Device support, Gboard version, and dictation language all matter. On supported Pixel configurations, automatic punctuation is associated with advanced or assistant voice-typing features.
If the setting is absent, or you use another keyboard
Check the available voice-typing options and update Gboard. An update does not add hardware- or language-restricted features to an unsupported configuration. Use spoken punctuation commands or a separate dictation service instead.
Also confirm that you are using Gboard. Samsung Keyboard and other Android keyboards have their own settings.
Windows 11: enable Auto-punctuation in voice typing
Path: Win + H > Settings gear > Auto-punctuation
- Click inside an editable text field.
- Press Windows key + H.
- Select the Settings gear in the voice-typing panel.
- Turn on Auto-punctuation.
- Start voice typing and speak naturally.
Windows voice typing operates at the operating-system level. It can enter text into compatible fields in email clients, text editors, and browser-based applications.
That distinction matters for Google Docs. Using Win + H inside a document is different from using Google Docs' own Tools > Voice typing command.
If Windows does not hear you
Check the selected microphone, microphone permissions, internet connection, and input language. Test in a simple text editor before diagnosing a problem with a particular website.
Microsoft Word: use Dictate settings
Path: Home > Dictate > Dictate toolbar > Settings gear
- Open a document in a version of Word that supports Dictate.
- Select Home > Dictate.
- Open the Settings gear in the dictation toolbar or pane.
- Select the correct spoken language.
- Enable Auto punctuation where available.
- Dictate a short statement and a question to check the output.
Word's dictation settings are separate from Windows voice typing. Changing one does not automatically change the other.
If Dictate or its settings are missing
The toolbar layout and feature availability differ between Word versions and platforms. If Dictate is missing, check your account or license, connected-experiences settings, and microphone permissions.
The Google Docs exception
Google Docs has no native auto-punctuation switch
Google Docs' native Tools > Voice typing feature does not provide automatic punctuation. Speaking a sentence naturally and waiting at the end does not substitute for saying "period".
For native Google Docs voice typing, use supported commands such as:
Want punctuation without speaking those commands? Use a different dictation engine, such as Windows voice typing, or create the text in a browser-based voice note tool and paste it into Docs.
You do not need to hunt through Google Docs preferences for a hidden toggle. There is no native auto-punctuation setting to turn on.
03 / Troubleshoot the words
Why Voice Typing Writes "Comma" Instead of a Punctuation Mark
Voice typing writes "comma" when it treats your command as ordinary speech rather than an instruction. If it writes "karma", that is a speech-recognition error; command support, language settings, microphone quality, and timing are the first things to check.
Fix the "spelling out words" glitch
Try these steps in order. Change one thing at a time so you can tell what helped.
-
Test a brief pause around the command
Say "Hello", pause for about half a second, say "comma", pause briefly again, then continue. This gives the engine a clearer boundary around the command. Treat this as a troubleshooting test, not a rule for every sentence. Long pauses can trigger unwanted sentence breaks in automatic mode.
-
Match the dictation language to your speech
Check the speech-recognition language, not just the keyboard layout. If your engine uses regional command vocabulary, try "period" for US English or "full stop" for UK English. Some tools recognize both.
-
Check whether that command is supported
An engine that supports "comma" might not support "colon" or "open parenthesis". Repeating an unsupported command more loudly will not add the feature.
-
Improve the microphone signal
Move away from background conversations, lower nearby music, and speak toward the microphone. If Bluetooth audio sounds muffled, test the device's built-in microphone. Compare the same sentence rather than changing several things at once.
-
Check which microphone button you activated
A browser app, an operating system, and a keyboard can each supply dictation. Their command lists and settings differ. Run one dictation engine at a time while testing.
-
Repeat the test in a plain text field
If punctuation works in a basic editor but not your target app, investigate that app's input handling. If it fails everywhere, restart dictation, verify the language, and check for software updates.
Testing comma one two three period
Testing, one two three.
If you get "Testing karma one two three period", focus on recognition quality and language settings before changing document formatting.
Why does dictation add two punctuation marks?
Some engines reconcile spoken commands with automatic punctuation. Others handle the combination poorly, especially around pauses.
If you see Hello,, or repeated sentence-ending marks, test the two modes separately: dictate naturally with auto-punctuation on, then try explicit commands with it off. Use whichever mode produces cleaner output in that application.
04 / Keep your train of thought
Why Speaking Punctuation Interrupts Your Thinking
Speaking punctuation makes you do two jobs at once: form an idea and describe its written structure. Automatic punctuation removes that extra task, but it still makes mistakes because pauses, sentence boundaries, and question intonation do not always match.
You know the feeling. You start explaining an idea, stop to say "comma", remember that you need a new paragraph, then lose the next point.
For a short text message, that interruption is manageable. For a research brainstorm, an email, or a meeting summary, it becomes a repeated switch between composing and formatting. You are no longer just speaking. You are editing aloud.
Say the marks
Spoken commands
BenefitYou choose where a supported punctuation mark goes.
Trade-offCommands interrupt your speaking rhythm, and unsupported commands may appear as words.
Let the engine decide
Automatic punctuation
BenefitYou can keep speaking in complete thoughts without naming each mark.
Trade-offThe engine can misread sentence boundaries, questions, and pauses. Review still matters.
Native auto-punctuation on iOS, Windows, and supported keyboards reduces that burden. It does not eliminate the need to review the result.
Common errors include:
- Misplaced commas: a hesitation becomes a comma even though the sentence does not need one.
- Run-on sentences: several related thoughts are joined without clear sentence boundaries.
- Incorrect question marks: a question receives a period, or an uncertain-sounding statement receives a question mark.
- Unhelpful paragraphing: a long explanation remains one block of text.
we can send the proposal tomorrow can finance approve the revised budget
We can send the proposal tomorrow. Can finance approve the revised budget?
The second sentence is a question because of its wording and structure, not simply because the speaker's pitch rises.
Why sentence structure matters, not just pauses
Natural language processing, or NLP, uses relationships between words to interpret sentence structure. That context helps distinguish a question from a statement even when the speaker's intonation is flat.
Built-in dictation also uses AI; the practical distinction is not "AI versus non-AI". It is quick text entry versus a fuller note-taking and transcription workflow.
05 / Speak first, review next
Dictate Naturally with SpeechText.AI's AI Voice Note Taker
For longer drafts, SpeechText.AI's AI Voice Note Taker provides browser-based dictation with automatic punctuation. Speak normally, review the text, and copy it into Google Docs, email, or another editor.
Cloud history and session recovery help you return to saved work. This is a practical alternative for Google Docs users who want to speak freely without adding "comma" and "period" to every thought: compose the note outside Docs, then transfer the punctuated text into your document.
No spoken punctuation commands are required.
Example: turn a fast brainstorm into a readable email
Imagine a student dictating an email between classes:
hi Maya I reviewed the project brief the scope looks good but the deadline conflicts with our exams can we move the review to Thursday I will send my notes tonight
Hi Maya, I reviewed the project brief. The scope looks good, but the deadline conflicts with our exams. Can we move the review to Thursday? I will send my notes tonight.
The words stay the same. The written structure becomes readable.
Automatic punctuation uses language context alongside speech cues such as pauses and cadence. That matters because "Can we move the review to Thursday?" should remain a question even when spoken quickly or with flat intonation.
The same workflow suits a professional dictating follow-up actions after a meeting. Speak the information while it is fresh, then inspect names, dates, commitments, and sentence boundaries before sharing it.
A simple browser-based workflow
- Open the AI Voice Note Taker. Allow microphone access and check that the browser is receiving the correct audio input.
- Speak in complete thoughts. Use your normal pace rather than inserting punctuation commands.
- Review the transcript. Check questions, proper names, numbers, and places where one idea should become two sentences.
- Check that the note is saved. Confirm that it is available in your history before leaving the page.
- Copy the text into its destination. Paste it into Google Docs, an email, or your preferred editor.
-
Speak naturally
can we meet Thursday I will send the agenda.
-
Recognize words
-
Add contextual punctuation
Can we meet Thursday? I will send the agenda.
-
Save to cloud history
-
Copy to Google Docs
What happens if you close the browser?
Ordinary phone voice typing usually puts text into whichever field is active. Whether that text survives depends on the destination app's draft-saving behavior.
SpeechText.AI's persistent cloud file history and session recovery provide a separate place to return to saved notes. Closing a browser window does not erase a note already stored in your history.
Saved is the important word. Cloud history protects stored material, not audio that was never captured or synced.
Before closing a long session: confirm that the note appears in your history. Keep a separate copy of critical text before making major edits.
06 / Start with a recording
How to Add Punctuation to Recorded Audio Files
Recorded speech does not need spoken punctuation commands: upload it to a transcription service that restores punctuation from speech and language context. SpeechText.AI supports this workflow for MP3, M4A, and WAV files, with capitalization, paragraph formatting, and speaker identification.
Voice memos, interviews, and meeting recordings
A meeting recording is different from live keyboard dictation. Nobody normally says "comma" during an interview, and there is no reason they should.
An engine that outputs only recognized words produces a wall of text. Many modern transcription tools also restore punctuation, so the absence of spoken commands is not the problem. The relevant question is whether the service processes the transcript into readable written speech.
SpeechText.AI combines automatic transcription with punctuation and formatting for uploaded recordings. Acoustic cadence and syntax help determine where sentences end, which clauses belong together, and where questions occur.
Turn a voice memo, interview, or meeting into punctuated text
- Upload the original recording. Use a supported file such as
.mp3,.m4a, or.wav. Prefer the original file over a repeatedly compressed copy. - Choose the correct language. A mismatch affects both recognized words and punctuation.
- Select an appropriate domain-specific model. Medical, legal, and technical terminology benefits from recognition settings suited to the subject.
- Configure speaker or channel processing as needed. Speaker diarization separates speaker turns. For recordings with speakers on separate channels, use multi-channel processing where supported.
- Process the recording. Automatic transcription adds words, punctuation, capitalization, and paragraph structure.
- Review important passages against the audio. Check names, dates, numbers, negations, and sentences whose meaning changes with punctuation.
- Save or export the reviewed transcript.
What each transcription feature does
| Feature | What it helps with | What still needs review |
|---|---|---|
| Automatic punctuation | Adds sentence endings, commas, and question marks | Ambiguous clauses and spoken fragments |
| Capitalization and paragraphing | Makes transcripts easier to read | Proper names and preferred editorial layout |
| Speaker diarization | Labels changes between speakers | Overlapping speech and assigning real names to labels |
| Domain-specific recognition | Improves handling of specialized vocabulary | Drug names, case references, technical identifiers, and other critical terms |
| Multi-channel processing | Preserves useful separation in recordings with distinct audio channels | Files where several voices already share one channel |
Scroll the table horizontally to compare all columns.
Speaker diarization is not the same as recognizing someone's identity. A label such as "Speaker 1" marks a voice; you still need to associate it with a person's name.
Why paragraph breaks also need a human check
Paragraph breaks involve interpretation. A recording does not contain an exact written layout waiting to be recovered. Review automatic paragraphing around speaker changes, topic shifts, and action items.
Fix the audio before blaming the punctuation
Punctuation quality depends partly on word recognition. If the engine mishears "can" as "can't", a neatly placed question mark will not repair the meaning.
Keep microphones close to speakers, reduce competing noise, and preserve separate speaker channels when the recording setup supports them. Converting a low-quality MP3 to WAV does not restore missing sound detail.
Working with sensitive recordings? For medical, legal, or client recordings, check consent, access, and retention requirements before uploading. Domain-specific recognition helps with vocabulary; it does not replace professional review.
