SDK Reference | BioLM Docs

Reference for the biolm Python package (pip install biolm-sdk). Product APIs below are the recommended entry points. For HTTP clients, legacy helpers, and deprecated imports, see biolm.core.

Product APIs

Inference

Model	Encode, predict, generate, and lookup with a bound model interface.
get_example, list_models	Model catalog browsing and copy-paste example generation (`biolm.models.examples`).

Protocols

Protocol	Load, validate, and inspect protocol YAML locally.
run_protocol()	Submit a protocol run and block until results are ready.
ProtocolClient	Submit, track, and download protocol runs programmatically.

Pipelines

GenerativePipeline	Multi-stage protein design with DuckDB caching and resumability.
Pipeline configs	`ScoringProtocolConfig`, `DirectGenerationConfig`, saturation mutagenesis, and related types.

Platform

Workspace	Workspace management (Python SDK upcoming).
Volume	Volume storage (Python SDK upcoming).
Finetune	XGBoost and DSM finetuning workflows.

Utilities

biolm.io	FASTA, CSV, JSON, and PDB file helpers.
biolm.hub	biolm-hub gateway discovery and configuration.

Core client (advanced)

Prefer product APIs above. Use biolm.core when you need direct HTTP control, async clients, or legacy imports.

biolm()	Legacy one-shot sync wrapper (re-exported from `biolm`).
BioLMApi	Sync HTTP client with schema access and manual batching.
BioLMApiClient	Async HTTP client; `await` its methods.
biolm.core.legacy	Deprecated; do not use in new code.

Plugins (optional)

Optional third-party backends under biolm.plugins. Install extras as needed (pip install biolm-sdk[mlflow]). Plugins are not imported from the top-level biolm package.

biolm.plugins.mlflow	Protocol result logging and MLflow-backed datasets.
protocol log	CLI: log protocol run results to MLflow.
dataset	CLI: list, upload, and download MLflow-backed datasets.

Full module index: biolm package.