How to transcribe audio in Word: step-by-step and limits

You recorded a meeting, an interview, or a class and need the text inside a Word document. There’s no reason to look for an external site: Word itself transcribes. The feature is called Transcribe, it lives in the Home menu, and it turns audio into text with speakers separated, ready to edit right away.

What Microsoft doesn’t write on that same screen are the rules of the game. Four accepted file formats, a 300-minute-per-month cap, and everything depending on OneDrive. This guide shows the complete step-by-step and, right after, what to do when one of those limits shows up in the middle of the work.

What Word’s Transcribe does

Transcribe converts speech to text and separates who said what. It works in two ways: recording directly in Word or receiving an existing audio file.

In practice, it delivers:

• Transcription with speakers identified one by one and fully editable labels.
• Audio with timestamps: click a passage and it plays exactly that part, with playback speed up to 2x.
• Two output buttons: add just a passage to the document or add the whole transcription.
• Support for 80+ languages, including Brazilian Portuguese.
• Recordings saved in the OneDrive Transcribed Files folder, which you can delete whenever you want.

When you use “Add all to document,” the text becomes regular Word content with a hyperlink to the original audio. That means you can share the file with your team and they can read it without needing the transcription panel open.

Step-by-step: transcribe an audio file in Word

1. Open the panel. In Word, go to Home > arrow next to Dictate > Transcribe. The item’s name changes with the Office language, but its location is always in the Home menu.
2. Choose the source. Under Upload audio, select the file from your computer. To record right now, use Start recording and speak normally; Word transcribes in the background while you keep working on the document.
3. Wait with the panel open. The transcription runs in the background and takes roughly as long as the audio’s duration. You can switch tabs or programs, but the panel must stay open.
4. Review speakers and passages. Each speech appears with a speaker label. Hover over the passage, use the edit icon to fix the text, and Change all Speaker to rename the person across all their speeches at once.
5. Send it to the document. Use Add section to document for a specific passage or Add all to document for the whole transcription. The text comes in formatted as a regular paragraph, and you edit it like any other content.

A detail that confuses many people: each document keeps a single transcription in the panel. If you start a new one, the previous one leaves the panel, but the passages already added to the document stay there.

On availability, Microsoft documents the feature in Word for the web and desktop Word on Microsoft 365 for Windows, with the feature fully documented for commercial tenants; in government tenants, transcription exists only in the web version.

The five limits that show up afterward

1. 300-minute monthly cap. Microsoft 365 subscribers transcribe up to 300 minutes of uploaded audio per month, the equivalent of 5 hours. Those with a Copilot license go up to 30,000 minutes. The cap is per account, not per document, and it resets at the start of each month.

2. Only four formats. The picker accepts .wav, .mp4, .m4a and .mp3. No .ogg, .opus, .flac or .webm. This has a practical consequence: a WhatsApp voice message, recorded in Opus, won’t go into the picker without prior conversion.

3. Everything goes through OneDrive. The uploaded or recorded audio is saved to the OneDrive Transcribed Files folder, and the panel must stay open during processing. On a restricted network or an account without OneDrive available, the flow stalls.

4. One document, one transcription. You can’t stack multiple transcriptions in the same file, nor process an entire folder of recordings at once.

5. The output is plain text. Word delivers the transcribed speech and nothing else: no summary, no task list, no mind map, no table of contents. You build all of that by hand afterward.

When Word is the best choice

Transcribe is worth using when you already pay for Microsoft 365, the audio is in a supported format, the month’s volume fits within 5 hours, and the end goal really is an editable Word document. For a one-off interview, a recorded class, or a short conversation, it’s the shortest path there is: no new signup, no export, no extra step.

When Word doesn’t solve it

The limit shows up in four recurring situations:

• Audio in a format the picker rejects. WhatsApp audio in Opus, an .ogg file recorded on Android, .flac from a recorder, a .webm video recorded in the browser.
• Volume above 5 hours per month. One week of daily one-hour meetings already consumes almost the entire cap.
• You need the result together with analysis. Executive summary, decisions, next steps, mind map: none of that comes out of Word’s panel.
• The material is a link, not a file. A recording on YouTube, a meeting saved in Meet, Zoom or Teams, a class on a course platform. Word only accepts files.

In these cases, the flow is to transcribe outside Word and bring the result into the document. The guide to transcribing audio with AI covers the available options, and the list of best apps for transcribing WhatsApp audio to text handles the most common rejected-format case.

How to generate the .docx without the 300-minute limit

In Sintesy, the path is short and starts wherever the file is:

1. Bring in the material. Paste the link (YouTube, Meet, Zoom, Teams, Teams Live) or upload the file from the dashboard. The import accepts mp3, wav, m4a, aac, ogg, oga, flac, opus, plus mp4, webm, mov, mkv, and also pdf, txt, doc and docx. The extensions Word rejects go straight in.
2. Let the synthesis run. Sintesy transcribes, identifies the speakers, and delivers a summary, searchable transcript and mind map on the same screen. It’s a product built for Brazilian Portuguese, so regional accents and technical terms don’t turn into noise in the middle of the text.
3. Download in Word. In the synthesis download menu, choose DOCX. The file arrives formatted, ready to edit, with no copy-and-paste step.

Two numbers matter before you start: the free plan unlocks 60 minutes of audio per day and 10 syntheses per day, and PDF and DOCX export is a Pro plan feature. That means you can test the whole flow for free and only unlock the download when the volume justifies it. The how much it costs to transcribe audio in 2026 comparison puts this model against the other two on the market.

Which path to choose

Quick questions

Does Word transcribe audio for free? Not standalone. Transcribe is part of Microsoft 365 and comes with 300 minutes of uploaded audio per month in the standard subscription.

How long does it take? About as long as the audio itself. A 40-minute file takes around 40 minutes to be ready.

Does Word accept WhatsApp voice messages? Not directly. WhatsApp records in Opus, a format that isn’t on the list of four accepted ones. Convert to .mp3 first, or use a tool that accepts .opus natively.

Does Word separate speakers? Yes, and the labels are editable, including renaming across all speeches at once.

Can it transcribe a full video? The .mp4 goes into the picker, but the limit is still the monthly cap and the supported formats. For a video that’s a link or a file outside the list, the path is a different flow.