How to Transcribe a Microsoft Teams Meeting

A Microsoft Teams meeting can produce a transcript three different ways, and which one you can actually use depends entirely on how your IT department configured your tenant. That's the frustrating part. One person on the call sees a tidy "Start transcription" button; the next person, on the same call, doesn't see it at all. Before you promise anyone meeting notes, it helps to know which path is open to you.

This guide covers all of it: the native Teams transcription feature, what to do when it's switched off, and how to turn a recording into clean text you can edit and share. Microsoft Teams transcription is reliable when it works, and there's a solid fallback when it doesn't.

What Teams offers on its own

Teams has genuinely good built-in transcription — when your organization allows it. There are two related features people mix up.

Live transcription produces real-time text during the meeting, attributed to speakers, and saves it afterward. Recording captures the audio and video to Microsoft's cloud (usually OneDrive or SharePoint) and can generate a transcript from that. They're separate toggles, and an admin can enable one without the other.

The speaker attribution in Teams is one of its stronger points because it ties text to signed-in accounts rather than guessing from voice alone. If everyone joins on their own device with their own login, you get reasonably accurate "who said what" labels straight out of the box.

How to start a transcript in Teams

If the feature is enabled for you, the steps are quick:

• In the meeting, click the three-dot More menu in the top control bar.
• Choose Record and transcribe, then Start transcription (or Start recording, which can include a transcript).
• A live transcript panel opens on the right. You can hide it without stopping it.
• When the meeting ends, the transcript saves and becomes available in the meeting chat and calendar event.

You can download the finished transcript as a DOCX or VTT file from the meeting details afterward. For the current specifics on where files land and which licenses include the feature, Microsoft's admin documentation is authoritative and it does change; the Microsoft Teams overview is a reasonable starting point for background.

When the button isn't there

Plenty of people never see the transcription option. Usual reasons:

• Your tenant admin disabled transcription or recording in the meeting policy.
• Your license tier doesn't include it.
• The meeting organizer's policy governs the call, not yours, so even if you could transcribe your own meetings, you can't transcribe theirs.
• Guest or external participants sometimes can't start it.

If any of those apply, native transcription is a dead end for that call. That's when you capture the meeting yourself and transcribe it separately.

Transcribing a Teams recording separately

The approach that doesn't depend on tenant policy: get a recording, then run it through a transcription tool. This is also worth doing when the native transcript exists but reads poorly — jargon-heavy calls and strong accents can leave the built-in version needing a lot of correction.

Getting a usable recording

You need a file. Options, in rough order of quality:

• A Teams cloud recording (if recording is allowed): download the MP4 from OneDrive or SharePoint.
• A local screen recorder: OBS or your operating system's screen capture will grab tab audio and video into an MP4.
• System-audio capture to MP3 or M4A: lighter and faster if you only need the words, not the video.

Always tell participants before you record. In many jurisdictions it's not optional, and even where it's legal, it's the decent thing to do. A quick verbal confirmation at the top of the call covers you.

Uploading to Transkio

With a file in hand, transcription is straightforward. Drop the MP4 into Transkio through the video-to-text tool and it pulls the speech out of the video directly — no need to extract the audio track yourself. If you already have an audio-only file like an MP3 or WAV, the audio-to-text path does the same job.

A few settings that change the result:

• Use the best recording you have. Muffled or heavily compressed audio drives up the error rate.
• Watch the file size. The free plan takes files up to 100 MB and recordings up to 30 minutes, which suits short calls; a long workshop may need a paid plan or splitting the file.
• Transkio supports 50+ languages, so a call in another language transcribes without extra fiddling.

AI-generated transcripts may contain errors — please review before relying on them.

Speaker labels: helpful, not flawless

If you need to know who said what and you're working from a recording rather than a native Teams transcript, speaker detection is a paid feature (Elite+). It separates voices by their acoustic signature, which works well when people sound distinct and take turns. It gets confused by similar voices, a shared conference-room mic, or people talking over each other. Read the labels as a strong draft and fix the mix-ups where they matter.

Cleaning up and exporting

No transcript comes out perfect. Give it a quick edit — correct proper nouns, fix any misheard figures, and trim filler if the document needs to read cleanly.

For output, the free plan exports TXT, SRT, VTT. Need a polished Word document to send around as official minutes? DOCX export is on paid plans (Pro+). And if the meeting was in one language but your audience needs another, translation is available on paid plans (Pro+); the audio-translation page explains what that covers and where machine translation still needs a human eye.

Picking the right method

Two clean decision points.

Stick with native Teams when

Your organization has transcription enabled, everyone's on their own signed-in device, and you mainly need a searchable, speaker-labeled record. It's automatic, the attribution is good, and the files download in usable formats. Spot-check the important lines and you're done.

Use a separate tool when

• Your tenant has the feature switched off.
• You're a guest on someone else's call and can't start it.
• The native transcript quality is poor and you want another pass.
• You need translation, subtitles, or a specific export the built-in feature doesn't produce.

The broader meeting transcription use-case page collects tips that apply no matter which platform your team uses.

A checklist you can reuse

Every Teams call I need notes from, I run the same loop:

1. Confirm recording out loud and get agreement.
2. Start the recording (native or a screen capture) before the discussion starts.
3. After the call, download the file and confirm it plays.
4. Upload it to Transkio and let it transcribe.
5. Skim the transcript, fixing names and numbers against memory.
6. Export in the format you need and share it.

For an hour-long meeting, that's about fifteen minutes of your time, most of it spent waiting on the tool.

The realistic takeaway

Microsoft Teams transcription is capable, but it's governed by policies you probably don't control, so build a habit that survives the days the button isn't there. Native transcripts when they're available; a captured recording plus a dedicated tool when they're not.

And keep the ceiling in mind. AI turns an hour of talk into text in a couple of minutes and nails clean, well-mic'd speech. It also mishears names, drops the thread during cross-talk, and can't tell that someone misspoke. The transcript is a fast, accurate-enough first draft — not a signed record. Your job shifts from transcribing to reviewing, which is far less work and produces notes people will actually trust.