Python SDK overview

The BioLM Python SDK lets you call BioLM models from Python with minimal setup: encode sequences, predict structures, and generate sequences. The recommended interface is Model; lower-level HTTP clients live in biolm.core.

Quick example

from biolm import Model

model = Model("esm2-8m")
result = model.encode(type="sequence", items="MSILVTRPSPAGEEL")

model = Model("esmfold")
result = model.predict(type="sequence", items=["MDNELE", "MENDEL"])

model = Model("progen2-oas")
result = model.generate(
    type="context",
    items="M",
    params={"temperature": 0.7, "num_samples": 2, "max_length": 17},
)

What you can do

• Encode sequences to get embeddings (e.g. ESM2-8M).

• Predict protein structures from sequences (e.g. ESMFold).

• Generate new sequences from context (e.g. ProGen2-OAS).

Ways to use the SDK

• Model inference: Use Model. See Client interfaces and biolm.models.

• Protocols, pipelines, platform: See SDK reference.

• Sync vs async: Model is sync; async apps use BioLMApiClient. See Concurrency.

• Batching, errors, rate limits, disk output: See Batching and Input Flexibility, Rate Limiting and Throttling, Error Handling.

• Advanced / legacy HTTP clients: See biolm.core.

Next steps

• biolm.models — Model interface and examples.

• biolm.core — biolm(), BioLMApi, and BioLMApiClient.

• biolm.pipeline — Pipeline config types and GenerativePipeline.

• Client interfaces — When to use which client interface.

• Frequently Asked Questions — Common questions.

• Full SDK module reference