> ## Documentation Index
> Fetch the complete documentation index at: https://bagel.softnanolab.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Getting Started

> Getting started with boileroom API

## Quick start

```python theme={null}
from boileroom import ESMFold

model = ESMFold(backend="modal")
result = model.fold("MKTVRQERLKSIVRI")

# Access the predicted structure
print(result.atom_array)  # list of Biotite AtomArray objects
```

## Available models

| Model     | Type                 | Method     | Description                                           |
| --------- | -------------------- | ---------- | ----------------------------------------------------- |
| `ESMFold` | Structure prediction | `.fold()`  | Meta's fast single-sequence structure prediction      |
| `ESM2`    | Embeddings           | `.embed()` | Protein language model embeddings (6 model sizes)     |
| `Chai1`   | Structure prediction | `.fold()`  | Diffusion-based structure prediction                  |
| `Boltz2`  | Structure prediction | `.fold()`  | Diffusion-based structure prediction with MSA support |

## Import patterns

```python theme={null}
from boileroom import ESMFold, ESM2, Chai1, Boltz2
```

## Constructor

All models share the same constructor signature:

```python theme={null}
Model(backend="modal", device=None, config=None)
```

<ResponseField name="backend" type="str" default="modal">
  Backend to use for execution. Supported values: `"modal"` (serverless GPU) or `"apptainer"` (local GPU via container). See [Backends](/boileroom-api/reference/backends).
</ResponseField>

<ResponseField name="device" type="str | None" default="None">
  GPU device identifier (e.g., `"cuda:0"`, `"cpu"`). If `None`, defaults to `"cuda:0"` when a GPU is available.
</ResponseField>

<ResponseField name="config" type="dict | None" default="None">
  Model-specific configuration overrides merged with defaults. Each model has its own config keys — see the model's page or [Configuration](/boileroom-api/reference/configuration).
</ResponseField>

## Context manager

All models can be used as context managers to ensure the backend is properly shut down:

```python theme={null}
with ESMFold(backend="modal") as model:
    result = model.fold("MKTVRQERLKSIVRI")
# Backend is automatically stopped
```

Without a context manager, the backend is cleaned up when the model instance is garbage collected or when the Python process exits.

## Folding models

<CardGroup cols={3}>
  <Card title="ESMFold" icon="dna" href="/boileroom-api/models/esmfold">
    Fast single-sequence protein structure prediction.
  </Card>

  <Card title="Chai-1" icon="atom" href="/boileroom-api/models/chai1">
    Diffusion-based structure prediction.
  </Card>

  <Card title="Boltz-2" icon="rocket" href="/boileroom-api/models/boltz2">
    Diffusion-based structure prediction with MSA support.
  </Card>
</CardGroup>

## Embedding models

<CardGroup cols={3}>
  <Card title="ESM2" icon="layer-group" href="/boileroom-api/models/esm2">
    Protein language model for sequence embeddings.
  </Card>
</CardGroup>

## Reference

<CardGroup cols={3}>
  <Card title="Output Types" icon="database" href="/boileroom-api/reference/output-types">
    Shared output types and cross-model comparison.
  </Card>

  <Card title="Configuration" icon="gear" href="/boileroom-api/reference/configuration">
    Static vs dynamic config, per-model config keys.
  </Card>

  <Card title="Backends" icon="server" href="/boileroom-api/reference/backends">
    Modal (serverless) and Apptainer (local) backends.
  </Card>
</CardGroup>