AIToday
Large Language ModelsSemantic Kernel BlogPublished: Aug 27, 2026, 01:01 JST1 min read

Microsoft Agent Framework adds channels for agents and workflows

Microsoft Agent Framework adds channels for agents and workflows

Key takeaway

  • Microsoft has added channels to Agent Framework, connecting agents to OpenAI Responses, Telegram, A2A, and MCP.

  • It uses a shared session model for cross-channel continuity.

  • Developers can integrate with their own web frameworks.

3 Key Points

  1. What happened

    Microsoft introduced channels in Microsoft Agent Framework for agents and workflows in Python. New packages connect agents to OpenAI Responses, Telegram, A2A, and MCP.

  2. Why it matters

    This lets developers expose the same agent or workflow through multiple interfaces without rewriting it. A shared session model can keep a conversation going across channels if the app maps identities to the same session ID.

  3. What to watch

    The team plans more channels and features. Developers can try the packages and samples, and share feedback via Python Channels issue #6265.

Ask the AI about this article →

Context & Analysis

This release addresses a common pain point for developers: integrating agents with the channels users already have. By providing protocol-specific packages, Microsoft aims to reduce duplication and maintenance. The shared hosting model lets applications own identity linking, authorization, and state storage, which is important for security and consistency. The emphasis on application-owned session mapping highlights that this framework is flexible but leaves key decisions to the developer. The call for feedback suggests Microsoft is iterating based on community needs, indicating room for growth in this area.

FAQ

What channels are supported?
The new packages support OpenAI Responses, Telegram, A2A, and MCP.
Can one agent work across multiple channels?
Yes, if the app maps different channel identities to the same session ID, a conversation can continue across channels.
Semantic Kernel BlogRead Original Article

Get the latest Large Language Models news every morning

For example, today's edition would include:

  • DataAgent launches with $10M to auto-fix Kubernetes faultsSiliconANGLE AI · 46m ago
  • SK Hynix custom HBM boosts inference up to 5.15xDIGITIMES Asia · 46m ago
  • Nvidia Earnings: Boring by Design, Avoiding a Consolidated WorldStratechery (Ben Thompson) · 46m ago

AI-summarized, only the topics you pick — one digest a day via Email, Slack, or Discord.

Free · takes 30 seconds · unsubscribe anytimeWhat is AIToday? →

Ask AI

Ask AI anything about this article. Q&As are published on this page for other readers too.

Related Articles

Next articleMCP roadmap: AI agent messaging, HTTP unification, identity