This repo contains code for running a production-grade multi-agent system on Railway. It's powered by Agno's AgentOS: a high-performance runtime for serving agents, multi-agent teams and agentic workflows.

This repo gives you:

• AgentOS runtime: for serving agents, multi-agent teams and agentic workflows.
• PostgreSQL database: for storing agent sessions, knowledge, and memories.
• A set of pre-built agents, teams and workflows to use as a starting point.

For more details, checkout Agno and give it a ⭐️

Follow these steps to run your AgentOS:

git clone https://github.com/agno-agi/agentos-railway.git
cd agentos-railway

We'll use OpenAI, Anthropic and Parallel Search services. Please export the following environment variables:

export OPENAI_API_KEY="YOUR_API_KEY_HERE"
export ANTHROPIC_API_KEY="YOUR_API_KEY_HERE"
export PARALLEL_API_KEY="YOUR_API_KEY_HERE"

We'll use docker to run the application locally and deploy it to Railway. Please install Docker Desktop if needed.

Run the application using docker compose:

docker compose up --build -d

This command builds the Docker image and starts the application:

• The AgentOS server, running on http://localhost:8000.
• The PostgreSQL database for storing agent sessions, knowledge, and memories, accessible on localhost:5432.

Once started, you can:

• View the AgentOS runtime at http://localhost:8000/docs.

• Open the AgentOS UI
• Login and add http://localhost:8000 as a new AgentOS. You can call it Local AgentOS (or any name you prefer).

The Agno Knowledge Agent is a great example of building a knowledge agent using Agentic RAG. It loads the Agno documentation into pgvector and answers questions from the docs.

To populate the knowledge base, run the following command:

docker exec -it agentos-railway-agent-os-1 python -m agents.agno_knowledge_agent

To load the data for the SQL Agent, run:

docker exec -it agentos-railway-agent-os-1 python -m agents.sql.load_f1_data

To populate the knowledge base, run:

docker exec -it agentos-railway-agent-os-1 python -m agents.sql.load_sql_knowledge

The Deep Knowledge Agent is a great example of building a deep knowledge agent using Agno.

To populate the knowledge base, run the following command:

docker exec -it agentos-railway-agent-os-1 python -m agents.deep_knowledge_agent

When you're done, stop the application using:

docker compose down

To deploy the application to Railway, run the following commands:

1. Install Railway CLI:

brew install railway

More information on how to install Railway CLI can be found here.

2. Login to Railway:

railway login

3. Deploy the application:

./scripts/railway_up.sh

This command will:

• Create a new Railway project.
• Deploy a PgVector database service to your Railway project.
• Build and deploy the docker image to your Railway project.
• Set environment variables in your AgentOS service.
• Create a new domain for your AgentOS service.

To add knowledge to the Agno Knowledge Agent, run the following command:

railway ssh --service agent_os

This command will open a ssh session to the AgentOS service.

Once you are in the ssh session, you can run the following command to add knowledge to the Agno Knowledge Agent:

python -m agents.agno_knowledge_agent

To update the application, run the following command:

railway up --service agent_os -d

This rebuilds and redeploys the Docker image to your Railway service.

To delete the application, run the following command:

railway down --service agent_os
railway down --service pgvector

Careful: This command will delete the AgentOS and PgVector database services from your Railway project.

To connect the AgentOS UI to the AgentOS server:

• Open the AgentOS UI
• Create a new AgentOS by clicking on the + button in the top left corner.
• Enter the AgentOS URL and click on the Connect button.
• You can add a local endpoint from your dev setup. To add the railway endpoint, you will be provided with a coupon code during the workshop.

Follow these steps to setup your local development environment:

We use uv for python environment and package management. Install it by following the the uv documentation or use the command below for unix-like systems:

curl -LsSf https://astral.sh/uv/install.sh | sh

Run the dev_setup.sh script. This will create a virtual environment and install project dependencies:

./scripts/dev_setup.sh

Activate the created virtual environment:

source .venv/bin/activate

(On Windows, the command might differ, e.g., .venv\Scripts\activate)

If you need to add or update python dependencies:

Add or update your desired Python package dependencies in the [dependencies] section of the pyproject.toml file.

The requirements.txt file is used to build the application image. After modifying pyproject.toml, regenerate requirements.txt using:

./scripts/generate_requirements.sh

To upgrade all existing dependencies to their latest compatible versions, run:

./scripts/generate_requirements.sh upgrade

Rebuild your Docker images to include the updated dependencies:

docker compose up -d --build

Need help, have a question, or want to connect with the community?

• 📚 Read the Agno Docs for more in-depth information.
• 💬 Chat with us on Discord for live discussions.
• ❓ Ask a question on Discourse for community support.
• 🐛 Report an Issue on GitHub if you find a bug or have a feature request.