AgentOS on Railway

Deploy AgentOS to Railway.

This template deploys AgentOS and PostgreSQL with pgvector to Railway. Automatic HTTPS and public domain included.

What’s Included

Component	Description
AgentOS	FastAPI server running your AI agents
PostgreSQL	Database with pgvector for embeddings
Automatic HTTPS	Railway handles TLS certificates
Public Domain	your-app.up.railway.app

Prerequisites

You’ll need:

• Docker Desktop
• uv (Python package manager)
• Railway account (free tier works)
• OpenAI API key

Install Railway CLI:

Mac

brew install railway

Deploy

Let’s start by creating the project

Create Your Project

1. Set up Python environment

   uv venv --python 3.12
   source .venv/bin/activate
   
2. Install Agno

   uv pip install -U 'agno[infra]'
   
3. Create from template

   ag infra create --template agentos-railway --name my-agentos
   cd my-agentos
   
4. Set your API key

   export OPENAI_API_KEY=sk-***
   

Included Agents

These agents are ready to use once deployed:

Pal

Personal assistant that learns from you. Stores notes, bookmarks, and research findings.

Knowledge Agent

RAG agent that answers questions from your documents. Pre-loaded with Agno docs.

MCP Agent

Connects to external tools via MCP. Extensible with any MCP-compatible service.

Deploy to Railway

1. Login and deploy

   railway login
   ./scripts/railway_up.sh
   

   This creates the project, provisions PostgreSQL with pgvector, and deploys your AgentOS. Takes ~2 minutes.

2. Get your domain

   railway open
   

   Find your URL in the dashboard (e.g., my-agentos.up.railway.app). Navigate to <railway-domain>/docs to see your AgentOS API.

Connect to Control Plane

1. Connect your AgentOS

   1. Go to os.agno.com
   2. Click “Connect OS” → Select “Live”
   3. Paste your Railway domain

   

Your AgentOS is live! Manage it from os.agno.com

Next Steps

Build Custom Agents

Create agents with your own tools and prompts.

Add Tools

Slack, GitHub, databases, and 100+ integrations.

Add Knowledge

Load documents, websites, and databases for RAG.

Add Interfaces

Deploy to Slack, Discord, WhatsApp, or expose via MCP.

Reference

Task	Command
Deploy updates	railway up --service agent_os -d
View logs	railway logs --service agent_os
Open dashboard	railway open
Set env variable	railway variables set KEY=value
Stop service	railway down --service agent_os
Scale replicas	Edit railway.json: {"deploy": {"numReplicas": 2}}
Switch model	railway variables set ANTHROPIC_API_KEY=sk-ant-***

Troubleshooting

'railway: command not found'

Install: brew install railway (Mac) or npm install -g @railway/cli

Deployment fails

Run railway init first, then ./scripts/railway_up.sh again.

Database timeout

PostgreSQL takes ~30s. Check: railway logs --service pgvector

502 error

Container starting. Wait 1-2 min. Check: railway logs --service agent_os