omicOS User Guide

A single-cell / spatial-omics analysis workbench —— powered by multi-agent LLMs, built on top of OmicVerse

What is this

omicOS is a workbench that brings together "multi-agent large language models" and "single-cell / spatial-omics analysis." You describe your analysis intent in natural language, and omicOS's team of agents automatically calls tools like OmicVerse and scanpy to produce real code, charts, and results inside a shared IPython kernel.

omicOS is made up of three product modules:

Part 4 is a cross-module collection of real analysis workflows, from data preparation and preprocessing through annotation and downstream analysis.

How to read this guide

• Part 1 shows you how to install omicOS on your own computer or server: installation, configuring environment variables, signing in to your cloud account, running remotely over SSH / HPC, and advanced usage of every command-line option. If you want to run an analysis, start here.
• Part 2 uses plenty of application screenshots to walk you through every feature of the app.omicos.cn web interface, step by step.
• Part 3 uses screenshots to explain how to configure and use the account and server-side management pages at auth.omicos.cn.
• Part 4 organizes practical use cases by analysis type. The Chinese edition begins with single-cell data preparation, preprocessing, and cell-type annotation.

📌 The guide is still being expanded. English placeholders remain available while the detailed use cases are prepared first in Chinese.

Quick start in three steps

# 1. Install omicos (via npm, the easiest way)
npm install -g @omicverse/omicos

# 2. Sign in to your cloud account
omicos login

# 3. Start the server and open the web interface
omicos serve
# the browser opens https://app.omicos.cn automatically

For the details of each step, other installation methods, and server scenarios, head to Part 1.