Thanks to visit codestin.com
Credit goes to docs.agno.com

Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.agno.com/llms.txt

Use this file to discover all available pages before exploring further.

Agent Runtime=Agent Server
AgentOS runs your agents as a service: reachable over an API, durable across restarts, scaled across replicas, and authenticated per request. 
from agno.agent import Agent
from agno.db.postgres import PostgresDb
from agno.models.openai import OpenAIResponses
from agno.os import AgentOS

db = PostgresDb(db_url="postgresql://user:pass@host:5432/agno")

agent = Agent(model=OpenAIResponses(id="gpt-5.5"), db=db)

agent_os = AgentOS(agents=[agent], db=db)
app = agent_os.get_app()

if __name__ == "__main__":
    agent_os.serve(app="my_app:app", reload=False)
This gives you a production-ready agent server. Persistent sessions, an HTTP API with streaming, JWT auth, tracing, and a scheduler, all from one stateless process.

​
What the runtime gives you

The runtime covers the ground between your agent code and a production service:
ConcernHow AgentOS handles it
HTTP APIAuto-generated endpoints for every registered agent, team, and workflow
StateSessions and memory persisted to your db
StreamingSSE on every run endpoint; tokens and tool calls stream as they happen
AuthJWT validation with RBAC scopes built-in
SchedulingIn-process cron that polls the database and fires due jobs
ObservabilityOpenTelemetry tracing into the same db, queryable with SQL
InterfacesSlack, Telegram, WhatsApp, A2A, AG-UI
Human in the loopPause runs for user confirmation, admin approval, or external execution
The same Agent(...) runs in a script, a test, or an AgentOS service. AgentOS is stateless: state lives in db, so you scale horizontally by adding replicas.
AgentOS can also serve agents built with the Claude Agent SDK, LangGraph, and DSPy. See Multi-framework support.

​
The primitives

ConcernPage
Run agents behind one HTTP APIServe as an API
Persist sessions, memory, knowledge, tracesStorage
Give the agent knowledge, dependencies, and live sourcesContext
Pause runs for confirmation or approvalHuman Approval
Trace every run and keep audit logsObservability
Authenticate, scope, and isolate every requestSecurity & Auth
Meet users in Slack, Telegram, WhatsApp, the browserInterfaces
Run recurring and multi-step workScheduling
Ship the service to productionDeploying

​
Explore

Serve as an API

The HTTP endpoints AgentOS generates and how to add your own.

Storage

Database backends and what gets stored.

Context

Knowledge, dependencies, and live data sources.

Human Approval

Pause runs for confirmation, approval, or external systems.

Observability

Tracing, run history, and audit logs in your own database.

Security & Auth

JWT validation, RBAC scopes, and per-request isolation.

Interfaces

Slack, Telegram, WhatsApp, A2A, and AG-UI.

Scheduling

In-process cron and multi-step workflows.

Deploying

Docker, Railway, AWS, GCP, and horizontal scaling.

​
Developer Resources