Can Microsoft Word transcribe audio files?

Yes. Word can transcribe an audio file you upload, but only with Transcribe. Transcribe sits on the Dictate menu, and it is a different feature from Dictate. Dictate types what you say as you speak. Transcribe takes a recording and returns text in a pane, and the transcript does not go into your document until you add it yourself ¹.

The more important question is whether your copy of Word has the feature at all. Microsoft limits it by platform and by account type.

Transcribe starts from the Dictate menu and takes an uploaded file

In Word, you start from Home, then Dictate, then Transcribe ². In the Transcribe pane, you select Upload audio and then choose a file in the file picker ³. Use it for a recording you already have. For speech typed into a document as you talk, Dictate is the right tool.

The platform rules narrow who can use it

Microsoft's note says the feature "is currently only available in Word for Microsoft 365 on Windows in Commercial Tenants" ⁴. Transcription for government tenants is available only in Word for the web ⁵.

Microsoft's support article also lists Word for the web and Word for Windows under "Applies To". Read that list together with the restriction note, which is the narrower statement. Microsoft's own page is the one to rely on, and you can check it directly before you plan around it.

Formats and the monthly limit decide whether a file fits

Transcribe currently supports .wav, .mp4, .m4a and .mp3 uploads ⁶. A Microsoft 365 subscriber can transcribe a maximum of 300 minutes of uploaded audio per month ⁷. A few long interviews can use up a whole month's allowance, so count the minutes before you start.

Time is the other limit. Microsoft says an uploaded file is transcribed in a time that depends on your internet speed, and that it can take up to about the length of the audio file ⁸. Keep the Transcribe pane open while it runs.

Speaker separation helps interviews, but it still needs a check

Microsoft says the Transcribe feature converts speech to a transcript "with each speaker individually separated" ⁹. For an interview, that saves real work. It does not remove the need to test the output. Listen to a two-minute stretch with several speakers, and check that each speaker change lands on the right person. The speaker diarization entry explains what the labels are.

It depends: when Word is enough and when it is not

If your organization needs proof that files stay private, check the data terms of your Microsoft 365 agreement. The transcription software guide explains the questions to ask any provider about where audio is processed.

What to do

1. Check your license and tenant. The feature is described for Microsoft 365 on Windows in commercial tenants.
2. Save the recording as .wav, .mp3, .m4a or .mp4. Convert anything else first.
3. Upload ten minutes of your hardest audio, including any crosstalk, and score the errors yourself. The word error rate entry explains how to count them.
4. Check the speaker labels, then copy the parts you need from the pane into your document.

If Word does not fit, compare Dragon, which can transcribe a file on Windows but where you correct another person's voice by hand, and Google Docs. For the Dictate side of Word, read where Dictate is in Word. For the wider set of tools, see the transcription overview.