Platform Overview

The BioLM platform brings together biological language models, workflow automation, and data management into a unified environment for molecular design and analysis.

It’s designed so researchers, developers, and teams can move seamlessly from individual model runs to complete reproducible experiments.

How the Platform Is Organized

At its core, BioLM connects four primary layers of functionality:

• Workspaces – The foundation of your BioLM experience. A workspace defines your environment, members, and compute resources. Everything you create — models, protocols, notebooks, and data — lives within a workspace.
• Models – Pretrained or fine-tuned biological language models that can predict, generate, or encode sequences. You can browse available models in the Console or build your own through finetuning.
• Protocols – Reproducible workflows that combine multiple models or steps, such as generating candidates, folding structures, and evaluating results. Protocols may include notebook sessions and persistent data storage.
• Data Assets – Your stored sequences, datasets, and results, accessible through Volumes, registries, and workspace resources.

Together, these layers form a structured workflow:

Workspaces provide organization → Models perform computations → Protocols orchestrate experiments → Data Assetscapture outputs.

How Users Interact With BioLM

BioLM can be accessed in multiple ways depending on your workflow:

• BioLM Console – The web interface for interactive exploration, running models, managing workspaces, and launching protocols.
• BioLM SDK – The Python interface for programmatic access, automation, and integration into external notebooks or pipelines.
• BioLM API – A REST interface for direct HTTP access to all platform features.
• BioLM Guides – Goal-oriented tutorials that demonstrate full workflows and real-world scientific use cases.

Each interface shares the same core architecture and data — what you create in one is visible in the others.

Typical Workflow

1. Start in a Workspace – Create or join a workspace and select an environment.
2. Choose or Train a Model – Run predictions, generate sequences, or fine-tune your own.
3. Build or Launch a Protocol – Combine steps into a reproducible workflow.
4. Review and Analyze Results – Access structured outputs through the Console or SDK.
5. Store or Share Data – Save notebooks and results to Volumes or registries for reuse.

This workflow is designed to be modular: each component can be used independently or as part of a full design loop.

Why It Matters

BioLM integrates tools that are usually fragmented across separate systems — modeling, analysis, and collaboration — into one cohesive platform.

By connecting these components, BioLM allows teams to move from hypothesis to validation faster, while maintaining transparency and reproducibility at every step.