Overview
Agno (formerly Phidata) is an open-source Python SDK for building AI agents and running them in production. You define an agent with a model, a set of tools, and instructions in plain Python, then run it as a service with tracing, scheduling, and access control built in.
It is aimed at Python developers and teams who want to own their agent stack instead of relying on a hosted black box. You keep control of your data, context, tools, memory, and human-review loops, store sessions and traces in your own database, and deploy on any platform that runs containers.
As a general agent framework, Agno covers both the build step and the runtime. Beyond the core agent loop, it ships a production API, 100+ tool integrations, context providers for sources like Slack and Drive, human-approval gates, and OpenTelemetry-based observability.
What it does
- Define agents in plain Python with a model, tools, instructions, and memory
- Run agents as a production API with 50+ endpoints, SSE, and websockets
- 100+ pre-built tool integrations plus context providers (Slack, Drive, MCP, custom)
- Human-approval gates to pause runs for confirmation and gate sensitive tools
- Built-in observability via OpenTelemetry tracing, run history, and audit logs
- Cron scheduling, background jobs, and container deploys to Docker, Railway, AWS, or GCP
Getting started
Install Agno with a model provider, then define and run an agent in a short Python script.
Install Agno
Install the agno package alongside a model provider client such as openai. The README's first-agent guide uses uv; plain pip also works.
uv pip install -U agno openaiSet your model provider key
Export the API key for the provider you chose so the agent can call the model.
export OPENAI_API_KEY=sk-...Build and run an agent
Create an Agent with a model and instructions, then stream a response with print_response.
from agno.agent import Agent
agent = Agent(
model="openai:gpt-5.5",
instructions="You are a helpful assistant.",
markdown=True,
)
agent.print_response("Say hello and tell me what you can do", stream=True)Commands and code are distilled from the project's own documentation — always check the official repo for the latest.
When to use it
- Build a coding or data assistant that lives in Slack and answers from your own context sources
- Stand up an internal agent platform with tracing, RBAC, and human approval that runs in your own cloud
- Expose an agent to users over Slack, Telegram, WhatsApp, Discord, or A2A interfaces
- Schedule recurring agent jobs (cron and background tasks) without extra infrastructure
How Agno compares
Agno alongside other open-source agent frameworks & builders tools AI/TLDR tracks, ranked by GitHub stars.
| Tool | Stars | What it does |
|---|---|---|
| AutoGPT | ★ 185k | One of the earliest autonomous agent projects, now a platform for building and running agents from reusable blocks and workflows. |
| Agno | ★ 40.8k | Python SDK for building, running, and managing AI agent platforms |
| AgentGPT | ★ 36.2k | AgentGPT lets you name a custom AI, give it a goal, and watch it plan tasks, run them, and learn from the results, all from a web browser. |
| LangGraph | ★ 35.2k | A library from the LangChain team for building stateful, graph-based agent workflows with explicit control over steps, memory, and human-in-the-loop checkpoints. |
| Composio | ★ 28.9k | Composio is an open-source SDK for Python and TypeScript that gives AI agents ready-made tools to act on real apps and APIs across many agent frameworks. |
| Semantic Kernel | ★ 28.2k | Microsoft's SDK for adding agents, plugins, and planning to apps across .NET, Python, and Java. |
| smolagents | ★ 27.9k | A minimal agent library from Hugging Face where the model writes and runs Python code to call tools and complete tasks. |
| Mastra | ★ 25.3k | A TypeScript framework for building AI agents and applications with workflows, RAG, memory, and observability built in. |