GitHub - microsoft/agent-framework: A framework for building, orchestrating and deploying AI agents and multi-agent workflows with support for Python and .NET.
2 hours ago
- #AI Agents
- #Multi-Language Framework
- #Graph Orchestration
- Microsoft's multi-language framework supports building, orchestrating, and deploying AI agents with .NET and Python implementations.
- Key features include graph-based workflows, AF Labs for experimental features, DevUI for development, and support for multiple LLM providers.
- Observability is integrated via OpenTelemetry, and there's a flexible middleware system for request/response processing.
- Quick start examples show how to create agents using Azure, OpenAI, and Microsoft Foundry to write haikus.
- Extensive documentation covers tutorials, user guides, migration from Semantic Kernel and AutoGen, and sample codes.
- Common issues include authentication errors with Azure CLI and API key problems, with tips on using credentials in production.
- Required environment variables for configuration include endpoints and deployment names for Azure OpenAI, Microsoft Foundry, and OpenAI.
- Security and compliance considerations are highlighted for using third-party servers or agents.