Behind IvoryOS: Empowering Scientists to Harness Self-Driving Labs for Accelerated Discovery
Published in Chemistry, Electrical & Electronic Engineering, and Materials

Why is it difficult to build and share a working SDL?
Open-source instrument drivers now exist for everything from pumps and heaters to balances and robotic arms. Python libraries support serial communication, workflow logic, and even AI-driven optimization. It seems like any lab with a few programmable tools and some scripting knowledge could build an SDL. But in practice, most labs still rely on manual workflows, because of the gap between individual libraries to an autonomous workflow.
Even when someone does write a working autonomous script, it often ends up trapped on one person’s laptop. Others in the same lab may not know:
- What the code does
- What parameters to change
- How to modify the code
This isn’t a failure of hardware or even programming skill. It’s a failure of code shareability.
How does IvoryOS make code shareable?
IvoryOS introduces two mechanisms that make SDLs easier to pass from one researcher to another:
1. Script Serialization
It automatically captures the current state of the lab script—what devices are loaded, what methods are available, and how they should be called. This “lab snapshot” can be updated without reconfiguration.
2. No-Code Workflow Interface
Users don’t need to know Python to run experiments. They can:
- Drag function blocks to a canvas
- Fill in input parameters with dropdowns or number fields
- Save results to variables
- Loop over experiments with either closed-loop experimentation or manual configuration
- Export the script for others to use or modify
This design minimizes onboarding time for new students and collaborators, especially those with less coding experience.
What makes IvoryOS a game-changer?
1. One-Line Installation
IvoryOS can be launched with a single line of code:
ivoryos.run(__name__)
No configuration files. No edits to your original Python script. You instantly get a fully interactive control panel and workflow builder, right in your browser.
2. Plug-and-Play across Labs
IvoryOS has been successfully deployed on six distinct self-driving labs across two institutions, each with different hardware, protocols, and development stages. It automatically adapts to any Python-based setup.
3. Built for Scale and Collaboration
Running as a web server, IvoryOS supports remote access out of the box. That means teams can design, monitor, or run experiments across labs or locations—no need to be physically present. Being browser-based means IvoryOS can seamlessly connect with other digital lab tools, dashboards, and cloud platforms.
Watch IvoryOS in action
In this video, we use a solubility platform at Telescope Innovations Corp. to perform color matching by mixing red and blue solutions to reach a target shade. Scientists define key functions—like mixing samples or analyzing color scores—but no longer need to write for-loops or optimization logic. These function iterations can be visually configured through the interface.
IvoryOS marks a pivotal step toward lab automation for all, empowering scientists to build and deploy their own SDLs across research domains. Our vision was made possible thanks to the support of the Acceleration Consortium at the University of Toronto, the Hein Lab at the University of British Columbia, and the amazing team at Telescope Innovations Corp.
What’s next?
We’re actively engaging with the open-source lab automation community to enhance IvoryOS’s compatibility with a wide range of lab instrument. Upcoming updates will introduce more modular and customizable deployment options. We’re also developing a Model Context Protocol (MCP) server—a natural language-based layer that enables more intuitive, conversational interaction with your Python scripts.
Try IvoryOS and contribute
IvoryOS source code is available at: https://gitlab.com/heingroup/ivoryos. Contributions are welcome!
Follow the Topic
-
Nature Communications
An open access, multidisciplinary journal dedicated to publishing high-quality research in all areas of the biological, health, physical, chemical and Earth sciences.
Related Collections
With collections, you can get published faster and increase your visibility.
Smart Materials for Bioengineering and Biomedicine
Publishing Model: Open Access
Deadline: Sep 30, 2025
Health in Africa
Publishing Model: Open Access
Deadline: Dec 31, 2025
Please sign in or register for FREE
If you are a registered user on Research Communities by Springer Nature, please sign in