CheMeleon
CheMeleon is a drug discovery & molecules tool whose code is published by JacksonBurns on GitHub, first public in 2026, described in Journal of Chemical Information and Modeling (2026). Low-data molecular property foundation model: a Chemprop D-MPNN pretrained on ~1M PubChem molecules to predict Mordred descriptors, then fine-tuned for potency, ADMET, solubility and other properties.
| Category | Drug discovery & molecules |
|---|---|
| Developer (GitHub) | JacksonBurns |
| Interface | Chemprop >=2.2.0 CLI/Python; fingerprint script |
| GitHub | JacksonBurns/chemeleon |
| Released | 2026-08-07 (source) |
| Paper | https://doi.org/10.1021/acs.jcim.6c01546 · Journal of Chemical Information and Modeling (2026) |
| Code license | MIT |
| Weights license | weights MIT on Zenodo |
| What this allows | Permissive (commercial OK) Use, modify and ship it inside closed-source commercial products; keep the license text and copyright notices. What do these mean? |
| License source | zenodo.org/api/records/15460715 · checked 2026-11-07 |
| Pricing | Open source |
| Tags | #molecular-property-prediction#foundation-model#admet#low-data#chemprop |
| Last checked | · links verified |
Can I use CheMeleon commercially?
Yes. The recorded license is permissive, so closed-source commercial use is allowed if you keep the notices.
- Code: MIT
- Weights: weights MIT on Zenodo
- Source: zenodo.org/api/records/15460715 · license checked 2026-11-07
What the license types mean · not legal advice.
Specs
| Task | — |
|---|---|
| Molecular representation | — |
| Open weights | — |
| GitHub stars | 167 |
| Interface | Chemprop >=2.2.0 CLI/Python; fingerprint script |
Compare with…
Pick 1–3 other Drug discovery & molecules tools to compare side by side with CheMeleon.
Related
Deep learning toolkit for drug discovery, materials and biology.
Open-source cheminformatics and machine learning toolkit.
Message passing neural networks for molecular property prediction.