Real-time Audio in Meetings Using a Bot

Question

Real-time Audio in Meetings Using a Bot

Anonymous

Using Bot, can I get audio in Real Time Meeting?

If yes, how can I get it. Thanks.

0 comments

Answer accepted by question author

Divyesh Govaerdhanan 11,395 MVP Volunteer Moderator

Hello,

Welcome to Microsoft Q&A,

Step 1: Use Microsoft Graph API for Teams Call Integration

Ensure your bot is correctly configured to capture media from Teams calls.
Reference: Calls and Meetings Bots Overview

Example: Using Microsoft Graph API to Join a Call

var

Step 2: Enable Real-Time Media Streaming

Configure the bot to extract real-time media using the Teams Bot Framework.
Reference: Real-Time Media Concepts

Example: Handling Audio Streams in a Bot

public

With this you will get the Audio stream of the team meeting, If you would like to convert this to text for transcription, You could use the Azure Speech service.

Step 3: Stream Audio to Azure Speech Service

Route the extracted media stream to Azure Speech Service for transcription.
Reference: Azure Speech-to-Text

Example: Sending Audio Stream to Speech-to-Text API

using

Step 4: Use Direct Line Speech for Enhanced Integration

If required, consider using Direct Line Speech to improve communication between the bot and Azure Speech Service.
Reference: Direct Line Speech Integration

Example: Enabling Direct Line Speech for a Bot

{
    "type": "directlinespeech",
    "serviceEndpoint": "https://directline.botframework.com/v3/directline"
}

As part of the next steps, to get your bot fully functional, deploy it using Azure Bot Service (Deploy a Teams Bot), ensure you grant the necessary Graph API permissions for calls and meetings (Graph API Permissions), and optimize audio processing for real-time transcription using Azure Speech Streaming API (Azure Speech Streaming API).

Please Upvote and accept the answer if it helps!

Shariq Majeed Khan 0 Reputation points

2026-02-12T15:22:24.38+00:00

Hi @Divyesh Govaerdhanan I have gone through the documentation you provided. I would appreciate some clarification on whether the referenced bot can be deployed in a serverless environment (for example, Azure Bot Service or Azure Web Apps), or whether it must be hosted on a dedicated server infrastructure such as a Virtual Machine.

3 additional answers

Your answer

Shariq Majeed Khan 0 Reputation points

2026-02-12T15:22:24.38+00:00

Hi @Divyesh Govaerdhanan I have gone through the documentation you provided. I would appreciate some clarification on whether the referenced bot can be deployed in a serverless environment (for example, Azure Bot Service or Azure Web Apps), or whether it must be hosted on a dedicated server infrastructure such as a Virtual Machine.

Answer 1

Noah Duncan 0

3 options:

Build an application-hosted media bot. Downside is your bot must run on Azure and be implemented using C#.
Use the Azure Communications Calling SDK with Teams interoperability. Downside is your bot will cause a notice to appear in the meeting saying the call has certain limitations.
Use browser automation to join the call as an anonymous user. Downside is it involves a large amount of development work. There is an open source project called Attendee that does the bulk of the required development work and offers real-time audio access.

0 comments

Answer 2

Jeeva 0

Can I capture live audio and video separately from a Microsoft Teams meeting in real time? If yes, are there any concepts or APIs that developers should be aware of, especially those that may be deprecated soon?

Mark Thomas 5 Reputation points

2025-09-16T17:11:49.62+00:00

Yes you can capture live audio and video separately. While I don't know about deprecation plans, as mentioned above, you'll need to familiarize yourself with the Graph API for the real time audio/video. If you need separate streams per speaker then that is where things get tricky. If you don't want to deal with deprecation issues at a later date, the 3rd party APIs are probably the easiest option since they'll need to figure out a way around any issues to support customers.

Answer 3

You can get real-time audio from a Microsoft Teams meeting building your own bot, but it requires setting up the full Graph API + Teams Calling Bot flow.

At a high level:

Register a Teams calling bot in Azure and enable it for media streaming. ◦ Requires the Calls.AccessMedia.All Graph API permission for app-hosted media.
Have the bot join the meeting as a participant via the Graph API.
◦ The bot receives the RTP audio packets in real time.
Route the audio into your transcription or processing service. ◦ You can use Azure Cognitive Services Speech-to-Text, or you can stream it to another service.

If you don’t want to handle the media pipeline, hosting, and permissions work yourself, there are APIs like Recall.ai’s Meeting Bot API that provide real-time audio, video, and transcripts from Teams (and other meeting platforms) through a single integration. That way you can skip the bot infrastructure and focus directly on what you want to build with the meeting data.

Feel free to DM me if you have questions