AgentOS is a production-ready FastAPI server that exposes your agents, teams, and workflows as REST APIs. It provides built-in endpoints for running AI components, managing sessions, and handling streaming responses.
Quick Start from agno import Agent from agno.api.os import AgentOS app = AgentOS( agents = [Agent( name = "assistant" , model = "gpt-4o" )] ) app.serve()
Constructor List of agents to expose via the API.
List of teams to expose via the API.
workflows
List[Workflow]
default: "[]"
List of workflows to expose via the API.
API title shown in documentation.
API description shown in documentation.
Host to bind the server to.
Port to run the server on.
Enable auto-reload for development.
Logging level: “debug”, “info”, “warning”, “error”.
Methods serve() Start the API server.
Parameters:
host (str): Override default hostport (int): Override default portreload (bool): Enable auto-reload
add_agent() Add an agent to the API after initialization.
app.add_agent(Agent( name = "helper" ))
add_team() Add a team to the API.
add_workflow() Add a workflow to the API.
app.add_workflow(workflow)
API Endpoints When you run AgentOS, the following endpoints are automatically created:
Agent Endpoints POST /agents//run Run an agent.
curl -X POST http://localhost:7777/agents/assistant/run \ -H "Content-Type: application/json" \ -d '{"message": "Hello!"}'
POST /agents//stream Run an agent with streaming.
GET /agents//sessions List sessions for an agent.
GET /agents//sessions/ Get a specific session.
Team Endpoints POST /teams//run Run a team.
POST /teams//stream Run a team with streaming.
Workflow Endpoints POST /workflows//run Run a workflow.
POST /workflows//stream Run a workflow with streaming.
Example Usage Basic Server
Multi-Component Server
With Database
Client Usage
from agno import Agent from agno.api.os import AgentOS from agno.models.openai import OpenAIChat assistant = Agent( name = "assistant" , model = OpenAIChat( id = "gpt-4o" ), instructions = "You are a helpful assistant." ) app = AgentOS( agents = [assistant], title = "My AI API" , version = "1.0.0" ) app.serve()
Deployment Docker FROM python:3.11-slim WORKDIR /app COPY requirements.txt . RUN pip install -r requirements.txt COPY . . CMD [ "python" , "app.py" ]
Production Settings from agno.api.os import AgentOS app = AgentOS( agents = [ ... ], host = "0.0.0.0" , port = 8000 , reload = False , log_level = "info" ) # Run with Gunicorn # gunicorn app:app -w 4 -k uvicorn.workers.UvicornWorker
