chai1

Predict protein, nucleic-acid, and small-molecule complex structures with Chai-1.

288|34|Updated Jul 6, 2026

Search Tags:#protein structure #complex analysis #protein folding #nucleic-acid #molecule prediction

Ready to run

npx skills add https://github.com/PKU-YuanGroup/OpenAI4S --skill chai1-pku-yuangroup

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: chai1
Source: https://github.com/PKU-YuanGroup/OpenAI4S/tree/main/skills/chai1
Command: npx skills add https://github.com/PKU-YuanGroup/OpenAI4S --skill chai1-pku-yuangroup

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) and assets (resource) components.

What problem does it solve?

Chai-1 is designed to address the challenges in protein, nucleic-acid, and small-molecule complex structure prediction, offering a solution for antibody-antigen or protein-ligand complex prediction and binder re-folding.

Core Features & Use Cases

Complex Structure Prediction: Accurately predict protein, nucleic-acid, and small-molecule complex structures.
Protein Folding: Re-fold designed binders as an alternative to AlphaFold for complex structure predictions.
Python Integration: Utilize Python for batched campaigns and GPU computation for structure predictions.

Quick Start

Run the run_inference function to predict the structure of a complex with a single FASTA file and GPU support.

Frequently Asked Questions about chai1

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I predict protein and nucleic-acid complex structures from a FASTA file?▼

To predict protein and nucleic-acid complex structures, run the `run_inference` function with a single FASTA file and GPU support. This executes batched structure predictions using the Chai-1 foundation model.

Can I use this model for antibody-antigen and protein-ligand complex prediction?▼

Yes, you can use this model for antibody-antigen and protein-ligand complex prediction. It accurately predicts protein, nucleic-acid, and small-molecule complex structures, also supporting binder re-folding.

Do I need a GPU for batch processing and protein folding analysis?▼

Yes, you need a GPU for efficient batch processing and protein folding analysis. The Python integration utilizes GPU computation to handle large-scale structure prediction campaigns effectively.

What is the best alternative to AlphaFold for complex structure prediction?▼

An alternative to AlphaFold for complex structure prediction is the Chai-1 foundation model. It re-folds designed binders and predicts complex protein, nucleic-acid, and small-molecule structures using Python and GPU computation.

Does Chai-1 support small-molecule and nucleic-acid structure prediction?▼

Yes, Chai-1 supports small-molecule and nucleic-acid structure prediction. It is specifically designed to address challenges in predicting complex structures involving proteins, nucleic acids, and small molecules.