How to build a Teams bot that streams and processes meeting audio?

Adrian Ćwiąkała 0 Punkty reputacji
2025-12-05T09:33:58.0166667+00:00

Hi all,

I’m exploring options for building a bot in Microsoft Teams that can:

join meetings or calls,

receive and process audio in real-time (e.g., transcription, AI processing),

  • ideally get separate audio streams per participant.

Most of the examples I’ve seen rely on the older .NET Calling/Media SDK (Application Hosted Media Bot), which looks a bit outdated and Windows-specific.

So I’m wondering:

Is there a modern approach to create a bot that can access meeting audio in real-time without using the legacy libraries like Microsoft.Skype.Bots.Media?

Can Bot Service give you:

  • the raw audio,
  • separate streams per participant,

Are there any limits I should know about, like latency, max participants, or concurrent calls?

Is there support for building such bots in Node.js, Python, or containerized/Linux environments, or is it still mainly .NET + Windows?

For production bots, is the Application Hosted Media Bot still the recommended approach, or is there a newer alternative?

I’d love to hear if anyone has successfully built a bot like this or has guidance on best practices and current recommendations.

Thanks!

Microsoft Teams | Microsoft Teams dla firm | Inne
Komentarze: 0

1 odpowiedź

Sortuj według: Najbardziej pomocne
Najbardziej pomocne Najnowsze Najstarsze
  1. Kudos-Ng 15,055 Punkty reputacji Pracownicy zewnętrzni firmy Microsoft Moderator
    2025-12-05T13:10:16.43+00:00

    Please note that this is the pl-pl forum. We kindly recommend posting your question in Polish so that more community members can assist you. Alternatively, you may consider posting in the English forum if you prefer to use your native language. We sincerely appreciate your understanding.

    Hi Adrian Ćwiąkała,

    Thank you for posting your question in the Microsoft Q&A forum. 

    To clarify upfront: I’m a forum moderator and don’t have the resources or environment to build and run a production media bot; what I can do is review and synthesize information that’s publicly available in Microsoft documentation and community discussions.

    In this case, I’ve taken time to review Microsoft’s official guidance, and for a native Teams bot that must access real‑time meeting audio/video, the recommended and production approach remains the Teams Real‑time Media Platform with an application‑hosted media bot built in .NET (using Microsoft.Graph.Communications.Calls.Media). This is the path that’s thoroughly documented and actively maintained for a bot joining Teams calls/meetings and receiving raw audio frames (20 ms, ~50 fps).

    Other languages (Node.js/Python) and environments (Linux) are not supported for native Teams real‑time media at this time.

    As for documented limits, Microsoft notes a bot can receive multiple main videos streams simultaneously up to 10 per media session and group calls scaling to roughly 350 participants, plus an operational constraint that each real‑time call is pinned to the VM that accepted it.

    References:

    I hope the insights above helpful.

    If the answer is helpful, please click "Accept Answer" and kindly upvote it. If you have extra questions about this answer, please click "Comment".

    Note: Please follow the steps in our documentation to enable e-mail notifications if you want to receive the related email notification for this thread.

    Czy ta odpowiedź była pomocna?

Twoja odpowiedź

Odpowiedzi mogą być oznaczone jako „Zaakceptowane” przez autora pytań i „Proponowane” przez moderatorów, co pomaga użytkownikom poznać odpowiedź rozwiązującą problem autora.