Microsoft.Agents.Extensions.Teams.AI

Microsoft.Agents.Extensions.Teams

Purpose

This is a Agents SDK based implementation of the former Teams AI library. Issues, releases, and source code available at https://github.com/microsoft/teams-agents

Changelog

Teams AI Agent APIs

Below is a detailed overview of the main APIs and their purposes:

Microsoft.Agents.Extensions.Teams.AI

This is the main library defined here. We leverage existing Agent SDK libraries to handle non-Teams and non-AI tasks.

• TeamsApplication: Represents a Teams application that can be used to build and deploy intelligent agents within Microsoft Teams.
• TeamsApplicationBuilder: A builder class for simplifying the creation of an Application instance.
• AI System: Responsible for moderating input and output, generating plans, and executing them. It includes components like the moderator, planner, and actions.
• ActionPlanner: A powerful planner that uses a Large Language Model (LLM) to generate plans. It can trigger parameterized actions and send text-based responses to the user.
• PromptManager: Manages prompt functions, data sources, and templates, allowing for the addition and retrieval of these components.
• GPTTokenizer: Implements a GPT Tokenizer for encoding and decoding text using the "gpt-4" model.
• JsonResponseValidator: Parses any JSON returned by the model and optionally verifies it against a JSON schema.

Microsoft.Agents.Hosting.AspNetCore

• AddAgentAspNetAuthentication: Configures authentication services based on provided settings in the configuration file, facilitating authentication and authorization for Microsoft Teams agents.

Microsoft.Agents.Authentication.Msal

• Authentication: Provides authentication capabilities using MSAL (Microsoft Authentication Library) for secure access to Microsoft services.

Microsoft.KernelMemory

• Kernel Memory: A library that allows indexing and querying any data using LLM and natural language, tracking sources, and showing citations.

Azure and OpenAI Integration

• Azure.AI.OpenAI: Provides integration with Azure OpenAI services for natural language processing.
• Azure.AI.ContentSafety: Ensures content safety by moderating input and output using Azure's Content Safety API.

Configuration and Deployment

• teamsapp.yml: Automates the creation of a Teams app registration, Azure AD app registration for a bot, and the deployment of infrastructure using ARM templates.
• appsettings.json: Contains configuration settings for token validation, connections, and logging.

Getting Started

We recommend getting started with the Teams Chef sample, available at

• Teams Chef Bot Sample - An AI-powered agent designed to assist developers in building Microsoft Teams apps.