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

# Install the bunker-fold Python Package

> Install bunker-fold, the official Python SDK for every Bunker Longevity model. Pick a package manager, verify the install, and run your first prediction.

`bunker-fold` is the official Python SDK for Bunker Longevity. A single package gives you a typed client for every model in the [catalog](/models/overview), including BunkerFold-3, GenoScan-LLM, MolGen-7, CryoNet, ToxPredict-v2, and EpiMark.

## Requirements

* Python 3.9 or later
* A Bunker Longevity API key ([Authentication](/authentication))

## Quick install

```bash theme={null}
pip install bunker-fold
```

## Pick a package manager

* [pip](/installation/pip): the default Python installer.
* [uv](/installation/uv): fast, resolver-first workflow for new projects.
* [Poetry](/installation/poetry): lockfile-driven dependency management.
* [Docker](/installation/docker): pre-built image with `bunker-fold` and CUDA runtime.

## Verify

```python theme={null}
import bunker_fold
print(bunker_fold.__version__)
```

## Configure your API key

Export your key as an environment variable so the SDK picks it up automatically:

```bash theme={null}
export BUNKER_API_KEY="sk_live_..."
```

Or pass it explicitly:

```python theme={null}
from bunker_fold import Bunker

client = Bunker(api_key="sk_live_...")
```

<Note>
  Never commit API keys to source control. See [Authentication](/authentication) for rotation and scope guidance.
</Note>

## First call

```python theme={null}
from bunker_fold import Bunker

client = Bunker()

result = client.models.bunkerfold_3.predict(
    sequence="MVLSPADKTNVKAAWGKVGAHAGEYGAEALERMFLSFPTTKTYFPHFDLSH"
)
print(result.confidence, result.structure_url)
```