ivoryos 1.5.19

an open-source Python package enabling Self-Driving Labs (SDLs) interoperability

IvoryOS: interoperable orchestrator for self-driving laboratories (SDLs)

A plug-and-play web interface for flexible, modular SDLs — you focus on developing protocols, IvoryOS handles the rest.

Join our community!

IvoryOS is an open-source project under active development. We welcome feedback, feature ideas, and contributions from anyone working on or interested in self-driving laboratories.

Join our Discord or Slack to ask questions, share use cases, and help shape IvoryOS.

---

Table of Contents

• What IvoryOS does
• System requirements
• Installation
• Features
• Demo
• Roadmap
• Contributing
• Acknowledgements

---

What IvoryOS Does

• Turns Python modules into UIs by dynamically inspecting your hardware APIs, functions, and workflows.
• Standardizes optimization inputs/outputs, making any optimizer plug-and-play.
• Provides a visual workflow builder for designing and running experiments.
• Adds natural-language control for creating and executing workflows, see IvoryOS MCP for more details.

---

System Requirements

Platforms: Compatible with Linux, macOS, and Windows (developed/tested on Windows).
Python:

• Recommended: Python ≥3.10
• Minimum: Python ≥3.7 (without Ax optimizer support)

Core Dependencies:

Click to expand

• bcrypt~=4.0
• Flask-Login~=0.6
• Flask-Session~=0.8
• Flask-SocketIO~=5.3
• Flask-SQLAlchemy~=3.1
• SQLAlchemy-Utils~=0.41
• Flask-WTF~=1.2
• python-dotenv==1.0.1
• pandas

Optional:

• ax-platform==1.1.2
• baybe==0.14.0
• nimo
• slack-sdk

---

Installation

From PyPI:

pip install ivoryos

32-bit Windows Installation

If you are deploying IvoryOS to a 32-bit Windows environment, pip may not find pre-compiled 32-bit binaries for certain modern packages (like greenlet, pandas, ``) on PyPI.

# 1. Install Flask-Session<0.7 (that doesn't depend on msgspec) and pandas (some pandas versions are not compatible with 32-bit Windows)
pip install Flask-Session<0.7
pip install pandas --user --only-binary=:all:

# 2. Install greenlet from conda-forge (32-bit PyPI wheels are no longer available)
conda install -c conda-forge greenlet -y

# 3. Finally, install IvoryOS
pip install ivoryos

Quick start

In your script, where you initialize or import your robot:

my_robot = Robot()

import ivoryos

ivoryos.run(__name__)

Then run the script and visit http://localhost:8000 in your browser. Use admin for both username and password, and start building workflows!

---

Features

Direct control:

direct function calling Devices tab

Workflows

• Design Editor: drag/add function to canvas in Design tab, use #parameter_name for dynamic parameters, click Prepare Run button to go to the execution configuration page
• Execution Config: configure iteration methods and parameters in Compile/Run tab.
• Design Library: manage workflow scripts in Library tab.
• Workflow Data: Execution records are in Data tab.

Logging

Add single or multiple loggers:

ivoryos.run(__name__, logger="logger name")
ivoryos.run(__name__, logger=["logger 1", "logger 2"])

Human-in-the-loop

Use pause in flow control to pause the workflow and send a notification with custom message handler(s). When run into pause, it will pause, send a message, and wait for human's response. Example of a Slack bot:

def slack_bot(msg: str = "Hi"):
    """
    a function that can be used as a notification handler function("msg")
    :param msg: message to send
    """
    from slack_sdk import WebClient

    slack_token = "your slack token"
    client = WebClient(token=slack_token)

    my_user_id = "your user id"  # replace with your actual Slack user ID

    client.chat_postMessage(channel=my_user_id, text=msg)

import ivoryos
ivoryos.run(__name__, notification_handler=slack_bot)

Use Input in flow control to get human input during workflow execution. Example:

Directory Structure

Created automatically in the same working directory on the first run:

click to see the data folder structure

• ivoryos_data/:
  • config_csv/: Batch configuration csv
  • pseudo_deck/: Offline deck .pkl
  • results/: Execution results
  • scripts/: Compiled workflows Python scripts
  • default.log: Application logs
  • ivoryos.db: Local database

---

Demo

Online demo at demo.ivoryos.ai. Local version in abstract_sdl.py

---

Roadmap

• Python property support (setter/getter)
• Support dataclass input
• Support **kwargs input
• dropdown input
• snapshot version control
• check batch-config file compatibility

---

Contributing

We welcome all contributions — from core improvements to new drivers, plugins, and real-world use cases. See CONTRIBUTING.md for details.

---

Citing

Click to see citations

If you find this project useful, please consider citing the following manuscript:

Zhang, W., Hao, L., Lai, V. et al. IvoryOS: an interoperable web interface for orchestrating Python-based self-driving laboratories. Nat Commun 16, 5182 (2025).

@article{zhang_et_al_2025,
  author       = {Wenyu Zhang and Lucy Hao and Veronica Lai and Ryan Corkery and Jacob Jessiman and Jiayu Zhang and Junliang Liu and Yusuke Sato and Maria Politi and Matthew E. Reish and Rebekah Greenwood and Noah Depner and Jiyoon Min and Rama El-khawaldeh and Paloma Prieto and Ekaterina Trushina and Jason E. Hein},
  title        = {{IvoryOS}: an interoperable web interface for orchestrating {Python-based} self-driving laboratories},
  journal      = {Nature Communications},
  year         = {2025},
  volume       = {16},
  number       = {1},
  pages        = {5182},
  doi          = {10.1038/s41467-025-60514-w},
  url          = {https://doi.org/10.1038/s41467-025-60514-w}
}

For an additional perspective related to the development of the tool, please see:

Zhang, W., Hein, J. Behind IvoryOS: Empowering Scientists to Harness Self-Driving Labs for Accelerated Discovery. Springer Nature Research Communities (2025).

@misc{zhang_hein_2025,
  author       = {Wenyu Zhang and Jason Hein},
  title        = {Behind {IvoryOS}: Empowering Scientists to Harness Self-Driving Labs for Accelerated Discovery},
  howpublished = {Springer Nature Research Communities},
  year         = {2025},
  month        = {Jun},
  day          = {18},
  url          = {https://communities.springernature.com/posts/behind-ivoryos-empowering-scientists-to-harness-self-driving-labs-for-accelerated-discovery}
}

---

Acknowledgements

Authors acknowledge Telescope Innovations Corp., UBC Hein Lab, and Acceleration Consortium members for their valuable suggestions and contributions.