Beyond Langchain : How Agno changed the way I build AI agents

I have been building agents for a while, and I noticed that most developers building AI agents today reach for LangChain by default, not because it's the best tool, but because it's the most visible one. This talk is about what happens when you look beyond the obvious.

In this talk, I will walk through Agno (formerly Phidata), an open-source framework for building multi-agentic systems.

To be very honest, it's actually a full platform split into two distinct layers:

1. The SDK -> This is the part you write code with. Pure Python, no special syntax to learn. It supports 30+ model providers, 100+ tool integrations, and 20+ vector databases, all behind one consistent API.

2. AgentOS ->This is the production layer. A FastAPI service that exposes 50+ ready-to-use endpoints, handles sessions, memory, tracing, RBAC, and human-in-the-loop flows.

The key takeaways from the talk would be:-

• What Agno is and how to actually use it to build your first agent?

• How it's different from tools like LangChain -> not a feature comparison, just another tool that makes building simpler.

• What it's capable of -> memory, tools, knowledge, multi-agent, and which ones you'd reach for first.