Skip to content
agent-os Docs
Brand App Docs

agent-os Documentation

What is agent-os?

Section titled “What is agent-os?”

agent-os is a SaaS observability dashboard for autonomous AI agent systems. Developers and small teams running LangGraph, CrewAI, or similar frameworks in production get a central control plane: real-time status, logs, errors, and security events — all in one place.

agent-os is not an agent framework. It’s an independent monitoring layer that works alongside existing systems.

Architecture

Section titled “Architecture”

The system consists of five main components:

ComponentTechnologyHosting
Dashboard AppLaravel 13 + Blade + SCSS + Reverb WebSocketRailway
Python SDKPython 3.10+, pip packagePyPI
JavaScript SDKTypeScript, npm packagenpm
Brand SiteNext.js 16, React 19, TypeScriptVercel
DocsAstro 6, StarlightCloudflare Pages

How it works

Section titled “How it works”
Your Agent (Python or JavaScript)
    │  pip install agent-os-sdk  /  npm install agent-os-sdk
    │  aos = AgentOS(api_key="...")
    │  graph = aos.instrument(graph)
    ▼
agent-os API (apps/app/)
    │  POST /api/ingest
    │  receives run events, errors, tool calls
    ▼
Postgres (persistent storage)
    │
    ▼
Dashboard (apps/app/)
    │  real-time updates via WebSocket (Laravel Reverb)
    └─► agent status, event stream, errors, costs, alerts

Quick Start

Section titled “Quick Start”

1. Create an account

Section titled “1. Create an account”

Sign up at agent-os.dev and grab your API key from Settings > API Keys.

2. Install the SDK

Section titled “2. Install the SDK”
Terminal window
pip install agent-os-sdk

Or for JavaScript/TypeScript:

Terminal window
npm install agent-os-sdk

3. Instrument your agent

Section titled “3. Instrument your agent”

Python (LangGraph):

from agent_os_sdk import AgentOS


aos = AgentOS(api_key="your-api-key")
graph = aos.instrument(compiled_graph, agent_name="my-agent")
result = graph.invoke({"messages": [...]})

JavaScript/TypeScript:

import { AgentOS } from "agent-os-sdk";


const aos = new AgentOS({ apiKey: "your-api-key" });
const result = await aos.trace("my-agent", async (runId) => {
  // your agent logic here
  return output;
});

From this point on, your agent sends events to agent-os automatically.

4. Open the dashboard

Section titled “4. Open the dashboard”

Log in and see your agent’s runs in real time: status, event stream, errors, costs, and tool calls.

Next steps

Section titled “Next steps”