agapi 9.10.2025

Python client and CLI for AtomGPT.org API (AGAPI)

AtomGPT.org API (AGAPI)

AGAPI provides examples of how to interact with https://atomgpt.org/

A significant amount of time in computational materials design is often spent on software installation, which creates a major barrier for newcomers to the field. AGAPI removes this layer by providing APIs that can be used directly for prediction and analysis, thereby lowering entry barriers and accelerating research.

Python SDK examples:

agapi/README.md

Prompt Examples

1. Access to Materials Databases

Prompt: List materials with Ga and As in JARVIS-DFT

Response:

2. Graph Neural Network Predictions

Prompt: Predict properties of this POSCAR using ALIGNN (Upload a POSCAR e.g. this file)

Response:

3. Graph Neural Network Force-field

Optimize structure from uploaded POSCAR file using ALIGNN-FF (Upload a POSCAR e.g. this file)

Response:

4. X-ray diffraction to atomic structure

Prompt: Convert XRD pattern to POSCAR (Upload a XRD file e.g. this file)

Response:

5. Live arXiv search

Prompt: Find papers on MgB2 in arXiv. State how many results did you find and show top 10 recent papers only.

Response:

References:

1. https://doi.org/10.1007/s40192-025-00410-9
2. https://doi.org/10.1016/j.commatsci.2025.114063
3. https://doi.org/10.1021/acs.jpclett.4c01126

Note: AGAPI (ἀγάπη) in Greek means unconditional love.