Pydantic AI vs Microsoft Agent Framework
Microsoft Agent Framework is the successor to AutoGen and Semantic Kernel, for Python, .NET and Go: agents, graph workflows with checkpointing, a configurable harness agent, and hosting on Microsoft Foundry and Azure Functions. Pydantic AI starts from one typed Agent on any model, with structured output validated by Pydantic, pydantic-graph for workflows, and capabilities that compose into complete agents such as Coder.
Pydantic AI is one part of a stack: the Harness SDK for capabilities and complete agents, Pydantic Evals for evaluation, Pydantic Graph for graph-based workflows, Pydantic Logfire for observability, and Pydantic itself for validation. The tables below cover the whole of it.
| Microsoft Agent Framework | Pydantic AI and Harness SDK | |
|---|---|---|
| Language | Python (also .NET and Go) | Python |
| License | MIT | MIT |
| Model providers | Many | Many |
| Extensibility | Tools, middleware, context providers | Capabilities and toolsets; 80+ with the Harness SDK |
| Harnesses | create_harness_agent, or build your own | Built-in Coder and Researcher, or compose your own |
| Observability | OpenTelemetry | OpenTelemetry, including Pydantic Logfire |
| Enterprise support | With a Microsoft Unified support agreement, when used with Azure AI services | For the framework itself, no platform required |
| Durable execution | Yes | Eight integrations |
| Interfaces | DevUI (sample app), AG-UI, A2A | CLI, web chat, AG-UI, Vercel AI, ACP (experimental), GitHub Agentic Workflows |
| Realtime voice | No | Realtime |
| Evals | Yes | Pydantic Evals |
| Image generation | Provider-hosted tools only | Image Generation |
| Microsoft Agent Framework | Pydantic AI and Harness SDK | |
|---|---|---|
| Multi-agent | Yes | Subagents, delegation, or pydantic-graph |
| Planning | Yes | Planning |
| Skills | Yes | Skills |
| Memory | Yes | Memory |
| Compaction | Yes | Compaction |
| Guardrails | Yes | Guardrails |
| Code sandboxes | Yes (beta) | Execution environments |
| Browser use | Computer-use tools; you run the browser | Web & research |
Can I use Pydantic AI with Microsoft Foundry? Yes. Run agents on Azure OpenAI and other Foundry model deployments, including Claude on Foundry, and voice agents on Azure AI Voice Live.
Can I build a coding agent on Pydantic AI? Yes. Give your agent the
Coder() capability, or assemble your own from the
same Agent; the harness repository has a
complete coding agent
built from the pieces Coder puts together.
Can I run my agents in CI? Yes. GitHub Agentic Workflows runs Pydantic AI agents
from a Markdown workflow file in GitHub Actions, or run a Python script directly with uv run;
nothing requires an Action.
Last checked: 2026-10-30