generated: '2026-07-18'
method: searched
source: https://github.com/chaidiscovery/chai-lab
cli:
  name: chai-lab
  package: chai_lab
  install:
  - pip install chai_lab==0.6.1
  - pip install git+https://github.com/chaidiscovery/chai-lab.git
  requirements: Linux, Python 3.10+, CUDA GPU with bfloat16 (A100/H100 80GB or L40S 48GB recommended)
  help: chai --help
commands:
- command: chai-lab fold
  summary: Predict a biomolecular structure from a FASTA input file, writing results to an output folder.
  usage: chai-lab fold input.fasta output_folder
  flags:
  - name: --use-msa-server
    summary: Generate MSAs via a remote ColabFold server for improved accuracy.
  - name: --use-templates-server
    summary: Use structural templates from a remote server.
  - name: --msa-server-url
    summary: Point at a custom ColabFold MSA server URL.
notes: >-
  First-party command-line interface shipped with the chai_lab Python package
  (the binary itself is catalogued in packages/). Added in chai_lab v0.4.3
  (2024-11-29); server URL configuration added in v0.4.4.