Run Locally

Run your agent platform locally.

Today we’re going to run an agent platform made of:

• AgentOS on FastAPI
• Postgres + PgVector

Incredibly simple, incredibly scalable. This stack goes a lot further than people realize.

Prerequisites

• Docker Desktop installed and running
• An OpenAI API key for models and embeddings

Run your agent platform

1. Clone the template

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

   To make this codebase yours, run rm -rf .git and push to your own git repo.

2. Configure your environment

   cp example.env .env
   

   Open .env and set OPENAI_API_KEY. Everything else has sensible defaults.

3. Start the platform

   docker compose up -d --build
   

   This runs two containers: a FastAPI app on port 8000 and a Postgres on port 5432.

4. Verify it's running

   Open http://localhost:8000/docs.

   You’ll see the OpenAPI spec. You can build incredibly powerful products on top of this.

   

You now have an agent platform made of AgentOS on FastAPI and Postgres. The AgentOS server exposes an large number of actions through a REST API.

AgentOS also comes with a UI at os.agno.com.

Connect the AgentOS UI

1. Open os.agno.com and log in.
2. Click Connect OS, choose Local, enter http://localhost:8000 as the URL, and name it Local AgentOS.
3. Click Connect.

You should see two agents:

Agent	Pattern	What it does
code_search	Context provider	Reads files in the workspace and answers questions about the codebase.
web_search	Direct tools	Searches the web and synthesizes answers grounded in citations.

Try a prompt against each:

“What’s in this repo?” → code_search lists the directory and explains the structure.

“What did Anthropic publish about agents recently?” → web_search returns a summary.

Open Sessions and Traces in the sidebar. Every run is captured with full message history, tool calls, and timing. This is what powers the iteration loop on the next page.

Summary

We now have a locally running agent platform with:

• Our agent runtime (AgentOS) running on port 8000 with request isolation, session management, scheduling, and 50+ endpoints.
• A Postgres database for storing sessions, memory, knowledge, and traces.
• A docs/ folder of Claude Code prompts for managing the entire agent development lifecycle: create, improve, extend, eval, and review.

Hot-reload is on. Edits to agents/*.py and app/*.py should be live in ~2s.

Next

Create an agent →