Overview
NIMO is developed by NIMS-DA and started as NIMS-OS (NIMS Orchestration System), the name the project keeps as its original identity. It treats artificial-intelligence methods and robotic experiments as interchangeable modules, so an exploration campaign can be assembled from the parts a laboratory already has.
The documented AI tools include Bayesian optimisation (PHYSBO), bound-less objective-free exploration (BLOX), phase diagram construction (PDC), probability that properties fall inside a target range (PTR), Bayesian optimisation for materials and process parameters (BOMP), combinatorial Bayesian optimisation (COMBI), nested Thompson sampling (NTS), an LLM-based experimental planner (LLMEP) and random exploration. Results can be followed with the bundled visualisation tools, and an IvoryOS interface is documented for designing and running workflows.
The README states that the package and its complete source code are distributed under the MIT License; the repository root held no separate LICENSE file when this entry was prepared. PyPI metadata (2.1.7) and the documentation site (3.0.0) are released separately, so pin the version you validate. The library proposes and records experiments; instrument drivers, calibration, safety limits and physical approval remain the operator's responsibility.
Key Features
- Closed-loop orchestration that couples AI exploration modules with robotic experiment modules.
- Documented exploration strategies ranging from Bayesian optimisation to objective-free exploration and phase diagram construction.
- IvoryOS interface for designing workflows, plus visualisation tools for exploring results in real time.
Use Cases
- Run an automated materials exploration campaign that alternates algorithmic proposals with robotic synthesis and measurement.
- Compare several exploration strategies on one candidate space before committing laboratory time.
How to Use
Install from PyPI with pip install nimo (Python 3.6 or newer, with physbo, scikit-learn, scipy, pyDOE3, matplotlib, numpy and anthropic as documented dependencies), or clone the repository and run pip install . inside it. The tutorial covers the candidates file, proposals produced by an exploration algorithm, generation of input files for robotic experiments, and analysis of the measured results.
Documentation: https://nims-da.github.io/nimo/en/
Sources and verification
This entry describes capabilities documented by the linked project sources. Catalog edits are not evidence of a successful run, compatibility test or scientific performance evaluation.
Read the editorial and verification policy