agent-framework and flatagents
About agent-framework
microsoft/agent-framework
A framework for building, orchestrating and deploying AI agents and multi-agent workflows with support for Python and .NET.
Supports graph-based workflows with streaming, checkpointing, and human-in-the-loop capabilities for deterministic function composition. Built-in OpenTelemetry observability, middleware system for request/response processing, and a DevUI for interactive agent development and debugging. Works with multiple LLM providers and includes experimental labs for benchmarking and reinforcement learning research.
About flatagents
memgrafter/flatagents
Flat, declarative agents and state machines for orchestrating LLMs.
Supports both single-agent tasks (FlatAgent) and multi-step orchestration via state machines (FlatMachine) with conditional branching, retry logic, and parallel execution—all defined in YAML with Jinja2 templating. Built around language-agnostic TypeScript schemas with lockstep versioning across Python and JavaScript SDKs, plus distributed worker patterns for scaling via registration and work backends. Integrates with multiple LLM providers (Anthropic, OpenAI, Cerebras) through centralized model profiles, with checkpoint/restore for fault tolerance and webhook hooks for remote execution.
Related comparisons
Scores updated daily from GitHub, PyPI, and npm data. How scores work