Excited to share Agno, a framework and runtime for multi-agent systems. Think of it as FastAPI for AI Agents.

At its core is the AgentOS, a high-performance server/runtime that helps you run and manage AI agents, multi-agent teams, and step-based agentic workflows — all inside your own cloud, with full privacy and no external data sharing.

What makes it different • Fast & lightweight — Agents instantiate in ~3μs and use ~6.6 KiB of memory on average (tested on M4 MacBook Pro). • Runtime architecture — Async, stateless, horizontally scalable runtime built on FastAPI. • Integrated UI — Test, monitor, and manage your agents and teams in real time. • Private by design — Runs entirely in your environment. No vendor lock-in, no telemetry, no external tracing.

A key distinction: the control plane connects directly to your AgentOS runtime from the browser — so no data is ever shared with 3rd party systems.

Docs & links • Docs → https://docs.agno.com • GitHub → https://github.com/agno-agi/agno • Examples → https://docs.agno.com/examples/introduction

What we'd love feedback on • Whether the architecture makes sense for you • Your thoughts on the DX and API • Use cases you'd apply this to

Happy to go deep on internals, performance, or multi-agent design patterns if there's interest.