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

# graspgenx

> Generate 6-DoF grasps for a named gripper with GraspGenX

Generate 6-DoF grasps for a named gripper with GraspGenX.

## Parameters

<ParamField body="aux_args" type="Dict[str, Any]">
  Must include "gripper\_name"; optional "num\_grasps", "planner" ("graspmoe"|"diffusion"), "grasp\_threshold", "camera\_extrinsics".
</ParamField>

<ParamField body="point_cloud" type="Union[str, np.ndarray, Sequence, None]">
  (N, 3) object point cloud.
</ParamField>

<ParamField body="depth_image" type="Union[str, Image.Image, np.ndarray, None]">
  Depth image; together with seg\_image and camera\_intrinsics, the alternative input to point\_cloud, converted to a point cloud server-side.
</ParamField>

<ParamField body="seg_image" type="Union[str, Image.Image, np.ndarray, None]">
  Segmentation mask selecting the object; part of the depth-based alternative to point\_cloud.
</ParamField>

<ParamField body="camera_intrinsics" type="Union[str, np.ndarray, None]">
  3x3 camera intrinsics matrix; part of the depth-based alternative to point\_cloud.
</ParamField>

<ParamField body="timeout" type="float | None">
  Optional HTTP timeout.
</ParamField>

## Returns

Dict\[str, Any]: Dict with:

* grasps: (N, 4, 4) array of grasp poses
* confidence: (N,) array of scores
* latency\_ms: server-reported latency

## Example

```python theme={null}
from grid_cortex_client import CortexClient, ModelType
import numpy as np
client = CortexClient()
pc = np.load("object_pc.npy")  # (N, 3)
out = client.run(ModelType.GRASPGENX, point_cloud=pc,
                  aux_args={"gripper_name": "robotiq_2f_85", "num_grasps": 64})
print(out["grasps"].shape)
```