NEST (Neural network Estimator of Stellar Times) is a python package designed to make the use of pre-trained neural networks for stellar age estimation easy.

It is based on Boin et al. 2026.

If you use NEST for your research, please acknowledge this by citing

@article{ Boin26,
    author = {{Boin, T.} and {Casamiquela, L.} and {Haywood, M.} and {Di Matteo, P.} and {Lebreton, Y.} and {Uddin, M.} and {Reese, D. R.}},
    title = {Stellar age determination using deep neural networks - Isochrone ages for 1.3 million stars, based on BaSTI, MIST, PARSEC, Dartmouth, and SYCLIST evolutionary grids},
    DOI= "10.1051/0004-6361/202558436",
    url= "https://doi.org/10.1051/0004-6361/202558436",
    journal = {A\&A},
    year = 2026,
    volume = 708,
    pages = "A215",
}

You can download the BibTeX citation file here: NEST.bib

With it, you can estimate the ages of stars based on their position in the Color-Magnitude Diagram and their metallicity. It contains a suite of Neural Networks trained on different stellar evolutionary grids. If observational uncertainties are provided, it can compute age uncertainties.

NEST Documentation

Installation :

• Using pip (preferred):

pip install astro-nest

• Download the source code of this repository, and run:

pip install .

• If you are looking for a quick way to test the Neural Networks, a web interface is also available here (click the image):

Dependencies:

• numpy
• scikit (optional but recommended for speed)
• tqdm (optional)
• matplotlib (optional)

The tutorial.ipynb notebook guides you through the package usage.