How to Transcribe Meetings Automatically (Zoom, Meet, Teams)

Meetings generate valuable information that's easy to forget. Automatic transcription captures every word, assigns speakers, and creates searchable records without manual effort. This article is part of our guide to speech-to-text use cases, covering how different professionals use transcription.

Whether you use Zoom, Google Meet, or Microsoft Teams, you have multiple options for automatic meeting transcription—from built-in platform features to specialized third-party tools.

Built-In Platform Transcription

Most video conferencing platforms now include native transcription features. While less feature-rich than dedicated tools, they require no additional setup or cost.

Microsoft Teams

Teams offers automatic recording and transcription that starts when your meeting begins:

1. Open Meeting Options before or during the meeting
2. Find "Record and transcribe automatically"
3. Set it to "Yes"

The transcript saves to the organizer's OneDrive for Business and appears in the meeting chat and Recap tab. Teams identifies speakers automatically and provides timestamps throughout.

Limitations: Transcription quality depends on audio clarity. Teams struggles with overlapping speakers and strong accents. The transcript format is basic—no AI summaries or action item extraction.

Google Meet

Google Meet's transcription integrates with your Google Workspace:

1. Create or edit a meeting
2. Click "Video call options"
3. Under "Meeting records," select "Transcribe the meeting"

Transcripts save to the organizer's Google Drive. Google Meet supports Gemini-powered note-taking that can generate summaries and action items from your transcript.

Limitations: Transcription is only available on certain Workspace plans. Chat messages aren't included in the transcript.

Zoom

Zoom's cloud transcription works with recorded meetings:

1. Enable "Audio transcript" in your account settings
2. Record your meeting to the cloud
3. Access the transcript in your recording page

Zoom provides speaker identification and searchable transcripts. You can edit transcripts and download them in multiple formats.

Limitations: Requires cloud recording (not local). Free accounts have limited cloud storage.

Third-Party Transcription Tools

Dedicated meeting transcription tools offer features beyond what platforms provide natively: AI-generated summaries, action item extraction, CRM integrations, and better accuracy.

Bot-Based Tools

These tools join your meeting as a participant and transcribe in real-time:

• Otter.ai connects to your calendar and automatically joins scheduled meetings. During the call, you see live transcription with speaker labels. After the meeting, you get summaries, timestamps, and action items. Otter reports up to 95% accuracy in optimal conditions.

• tl;dv offers a free tier with unlimited meeting recording and transcription. It generates AI-powered summaries and integrates with Zoom, Google Meet, and Teams.

• Fathom provides unlimited free transcription for individuals and syncs summaries directly to CRMs like Salesforce and HubSpot.

Trade-off: Participants see when a bot joins the meeting. Some people feel uncomfortable speaking freely with a visible transcription bot present.

Bot-Free Tools

These tools capture audio directly from your device without joining the meeting:

• Jamie records locally on your device and transcribes without any bot joining the call. It works even offline—useful in cafes or with unreliable internet. Jamie is GDPR-compliant and deletes recordings after transcription.

• Krisp captures audio at the system level and transcribes any meeting from any platform. It offers both recording and transcript-only options, with strong privacy controls.

• Bliro taps your device's direct audio source to transcribe. It works on desktop and mobile, independent of browsers or specific platforms.

Trade-off: Bot-free tools may have slightly lower accuracy since they don't access the meeting's speaker-separated audio streams.

Speaker Diarization: Who Said What

Accurate speaker identification transforms a wall of text into a useful record. Here's what affects diarization quality:

What helps accuracy:

• Clear audio with minimal background noise
• Speakers taking turns rather than overlapping
• 2-10 participants (fewer is better)
• Distinct voices among speakers

What hurts accuracy:

• Multiple people talking simultaneously
• Poor microphone quality or echo
• Large meetings with many speakers
• Similar-sounding voices

Some enterprise tools pull speaker data directly from the meeting platform, achieving near-perfect speaker identification with actual names. Consumer tools rely on voice analysis, which works well for small meetings but degrades with more participants.

Overlapping speech remains the biggest challenge. When two people talk at once, most transcription tools either miss content or attribute it to the wrong speaker.

Offline and Local Transcription

Privacy-conscious organizations often prefer transcription that doesn't send audio to external servers. Several approaches exist:

Local device transcription: Tools like Jamie process audio entirely on your device. The audio never leaves your computer, and recordings delete after transcription completes.

Self-hosted solutions: Open-source tools like WhisperX combine OpenAI's Whisper model with speaker diarization. You run everything on your own infrastructure with no data leaving your network.

Meeting recordings: Download your Zoom or Teams recording, then transcribe it using any local tool. This adds a step but gives you full control over your data.

For most users, the privacy controls offered by reputable third-party tools are sufficient. Enterprise environments with strict compliance requirements may need local or self-hosted solutions.

Choosing the Right Approach

Your choice depends on three factors:

Budget: Built-in platform transcription is free but basic. Third-party tools range from free tiers (tl;dv, Fathom) to $15-30/month for premium features.

Features needed: If you just want a searchable record, platform transcription works. If you need AI summaries, action items, or CRM integration, third-party tools deliver more value.

Privacy requirements: Bot-free tools offer discretion. Local processing keeps audio off external servers. Enterprise tools provide compliance certifications.

For occasional transcription needs, lightweight tools offer a middle ground. You upload recordings after the meeting and get transcripts with speaker identification—no subscription required, no bot joining your calls. This works well for professionals who don't need every meeting transcribed but want quality transcription when it matters.

Conclusion

Automatic meeting transcription is now accessible across every major platform. Start with your existing tools—Zoom, Meet, or Teams all include basic transcription. When you need better accuracy, speaker identification, or AI-powered summaries, third-party tools fill the gap.

For professionals who transcribe occasionally, tools like Scriby provide pay-as-you-go transcription with speaker diarization. Upload your meeting recording, get an accurate transcript with speaker labels, and pay only for what you use—no monthly subscription or bot joining your calls.