biolm hub | BioLM Docs

Connect the CLI and SDK to a running biolm-hub gateway (bh serve locally or a deployed Modal gateway). Platform login and protocols still use biolm.ai; only model inference is routed through the hub.

Prerequisites. Deploy and serve models from the biolm-hub repository:

bash

git clone https://github.com/BioLM/biolm-hub
cd biolm-hub
make install
source .venv/bin/activate
bh deploy esm2
bh serve

Typical workflow:

bash

bh serve
biolm hub set
biolm login
biolm model run esm2-8m encode -i seq.json

By default biolm hub set points at http://127.0.0.1:8000 and saves hub_api_url to ~/.biolm/config.yaml. BIOLM_BASE_API_URL in the environment takes precedence over the saved config. Browse models in the browser at http://127.0.0.1:8000/catalog while bh serve is running.

Route model inference through a local or self-hosted biolm-hub gateway.

Use biolm hub set after bh serve (or pointing at a deployed gateway). Platform login and protocol commands still target biolm.ai unless overridden in the environment.

Usage

bash

biolm hub [OPTIONS] COMMAND [ARGS]...

biolm hub set

Save a biolm-hub gateway URL so biolm model uses local or self-hosted inference.

Writes hub_api_url to ~/.biolm/config.yaml. Platform login and protocols still use biolm.ai unless BIOLM_BASE_DOMAIN is set.

Usage

bash

biolm hub set [OPTIONS] [URL]

Arguments

URL: Optional argument

biolm hub status

Show saved hub configuration, the active model API URL, and live gateway health.

Usage

bash

biolm hub status [OPTIONS]

biolm hub unset

Remove saved hub settings and revert model inference to the hosted biolm.ai API.

Has no effect on BIOLM_BASE_API_URL when that environment variable is set.

Usage

bash

biolm hub unset [OPTIONS]