> ## 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.

# CryoNet: Cryo-EM 3D Reconstruction Model

> CryoNet reconstructs 3D density maps from single-particle cryo-EM micrographs at near-atomic resolution using 10x fewer particles than standard pipelines.

CryoNet is a 540M parameter cryo-EM reconstruction model. Given a set of single-particle micrographs, it produces a 3D density map at near-atomic resolution while using roughly 10x fewer particles than conventional pipelines.

## At a glance

| Field      | Value                                       |
| ---------- | ------------------------------------------- |
| Category   | Imaging                                     |
| Parameters | 540M                                        |
| Version    | v1.4                                        |
| Endpoint   | `POST /v1/models/cryonet/predict`           |
| Output     | 3D density map plus FSC and symmetry report |

## Install

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

See [Installation](/installation) for full setup, including `uv` and `poetry`.

## When to use

* Reconstruct a 3D map from a small particle set.
* Get a fast preliminary map during a data collection session.
* Resolve symmetry ambiguities before running a full refinement.

## Example

<CodeGroup>
  ```bash cURL theme={null}
  curl -X POST https://api.bunkerbio.com/v1/models/cryonet/predict \
    -H "Authorization: Bearer $BUNKER_API_KEY" \
    -H "Content-Type: application/json" \
    -d '{
      "particles": 12400,
      "pixel_size": 1.06,
      "voltage": 300,
      "defocus_range": [1.2, 3.0],
      "input_url": "s3://my-bucket/particles.star"
    }'
  ```

  ```python Python theme={null}
  import requests, os

  r = requests.post(
      "https://api.bunkerbio.com/v1/models/cryonet/predict",
      headers={"Authorization": f"Bearer {os.environ['BUNKER_API_KEY']}"},
      json={
          "particles": 12400,
          "pixel_size": 1.06,
          "voltage": 300,
          "defocus_range": [1.2, 3.0],
          "input_url": "s3://my-bucket/particles.star",
      },
  )
  print(r.json())
  ```
</CodeGroup>

```json Response theme={null}
{
  "resolution_angstrom": 2.8,
  "fsc_threshold": 0.143,
  "symmetry": "C1",
  "map_size": [256, 256, 256],
  "processing_minutes": 14,
  "particles_used": 11820,
  "map_url": "https://cdn.bunkerbio.com/maps/xyz789.mrc"
}
```

## Parameters

<ParamField path="input_url" type="string" required>
  Signed URL or S3 path to a particle STAR file or a directory of micrographs.
</ParamField>

<ParamField path="particles" type="integer" required>
  Number of particles in the input set.
</ParamField>

<ParamField path="pixel_size" type="number" required>
  Pixel size in angstroms.
</ParamField>

<ParamField path="voltage" type="integer">
  Accelerating voltage in kV. Typical values are `200` or `300`.
</ParamField>

<ParamField path="defocus_range" type="number[]">
  Two-element array `[min, max]` in micrometers.
</ParamField>

<ParamField path="symmetry" type="string">
  Optional symmetry hint (`C1`, `C2`, `D2`, `I`, etc.). Defaults to automatic detection.
</ParamField>

## Response fields

<ResponseField name="resolution_angstrom" type="number">
  Reconstructed resolution in angstroms at the reported FSC threshold.
</ResponseField>

<ResponseField name="symmetry" type="string">
  Detected or applied symmetry group.
</ResponseField>

<ResponseField name="particles_used" type="integer">
  Number of particles that contributed to the final map after outlier rejection.
</ResponseField>

<ResponseField name="map_url" type="string">
  Signed URL to the reconstructed density map in MRC format. Valid for 24 hours.
</ResponseField>