If you're wondering how to get started with Foundry or want to see it in action, you're in the right place!
| # | Name | Time | Description |
|---|---|---|---|
| 00 | Hello Foundry | 5 min | Absolute basics - your first dataset |
| 01 | Quickstart | 5 min | Search, load, use in ML workflow |
| 02 | Working with Data | 15 min | Schemas, splits, PyTorch/TensorFlow |
| 03 | Advanced Workflows | 20 min | Publishing, HuggingFace, CLI, MCP |
Each folder contains a notebook and requirements.txt file. The notebooks can be run locally or in Google Colab.
| Example | Domain | Description |
|---|---|---|
| bandgap | Materials | Band gap prediction |
| oqmd | Materials | Open Quantum Materials Database |
| zeolite | Chemistry | Zeolite structure analysis |
| dendrite-segmentation | Imaging | Microscopy segmentation |
| atom-position-finding | Imaging | Atom localization |
from foundry import Foundry
f = Foundry() # HTTPS download by default
results = f.search("band gap", limit=5)
dataset = results.iloc[0].FoundryDataset
X, y = dataset.get_as_dict()['train']Cloud environments (Colab, etc.):
f = Foundry(no_browser=True, no_local_server=True)For large datasets with Globus:
f = Foundry(use_globus=True) # Requires Globus Connect PersonalCLI:
foundry search "band gap"
foundry schema <doi>If you have any trouble, check our documentation or create an issue.