> ## 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.

# ESMFold

> Object that uses ESMFold to predict structure of proteins from sequence.

`ESMFold` is a folding oracle that wraps Meta's ESMFold model for single-sequence protein structure prediction. It returns 3D atomic coordinates along with confidence metrics (pLDDT, pTM, PAE) that are consumed by structure-based energy terms.

For multimer prediction, ESMFold uses either a glycine linker between chains or a gap in positional embeddings (default: skip of 512) to signal chain boundaries. These can be configured via the `config` parameter.

<Info>
  ESMFold inference is powered by [boileroom](/boileroom-api/models/esmfold). boileroom handles model loading, GPU execution, and dependency isolation — either serverlessly via Modal or locally via Apptainer. See the [boileroom ESMFold reference](/boileroom-api/models/esmfold) for backend configuration details.
</Info>

## Parameters

<ResponseField name="use_modal" type="bool" default="False">
  Whether to run ESMFold on Modal's serverless GPU infrastructure. Set to `True` for serverless execution (no local GPU required), or `False` for local GPU execution.
</ResponseField>

<ResponseField name="config" type="dict[str, Any]" default="{}">
  Model-specific configuration. Common keys:

  * `glycine_linker` (str): linker sequence between chains (e.g., `"GGGG"`). Default: no linker.
  * `position_ids_skip` (int): gap in positional embeddings between chains. Default: `512`.
</ResponseField>

<ResponseField name="modal_app_context" type="App | None" default="None">
  Optional Modal app context for reusing an existing Modal session.
</ResponseField>

## Attributes

<ResponseField name="result_class" type="Type[ESMFoldResult]">
  The result class for this oracle. `ESMFoldResult` extends `FoldingResult` with ESMFold-specific attributes (pLDDT, pTM, PAE arrays).
</ResponseField>

## Methods

### fold

Fold a list of chains using ESMFold.

**Parameters**

<ResponseField name="chains" type="List[Chain]" required>
  The chains to fold. For multimer prediction, pass multiple chains.
</ResponseField>

## Example

```python theme={null}
import bagel as bg

# Create an ESMFold oracle using Modal
esmfold = bg.oracles.ESMFold(use_modal=True, config={
    "glycine_linker": "GGGG",
    "position_ids_skip": 1024,
})

# Use with energy terms
state = bg.State(
    chains=[binder, target],
    energy_terms=[
        bg.energies.PTMEnergy(oracle=esmfold, weight=1.0),
        bg.energies.PLDDTEnergy(oracle=esmfold, weight=1.0, residues=[binder_residues]),
    ],
    name="binding",
)
```