AI Agent Hub
Back to plugins
🤖

xby-pose

Model Inference Updated 2026.09.02

Run the following command in DeepSeek Harness:

dsh plugin install xby-skill/xby-pose

Paste the following prompt into your AI chat to install this plugin:

Run dsh plugin install xby-skill/xby-pose in DeepSeek Harness to install this plugin; the source code is available at https://github.com/xby-skill/xby-pose .

About this plugin

Integrating human pose estimation into a DSH workflow typically means wiring up inference pipelines and handling multiple input formats on your own. xby-pose packages all of that into a drop-in tool so you can focus on application logic instead of low-level engineering.

Powered by the YOLO26 model under the hood, it detects every person in an image and returns a bounding box plus 17 keypoint coordinates covering the nose, both eyes, both ears, both shoulders, both elbows, both wrists, both hips, both knees, and both ankles. Three input modes are supported — image URL, BASE64-encoded data, and local file path — making it straightforward to slot into diverse pipelines, chat sessions, and automation scripts.

It is a natural fit for developers and teams who need fast, in-context pose recognition for motion analysis, fitness posture checking, virtual try-on pre-processing, or any workflow where a quick structured readout of a person is the next step. Grab an API key, call the tool inside DSH, and ship without maintaining a GPU cluster.

Use Cases

  • Read person pose and keypoints right inside a DSH chat session
  • Feed structured 17-keypoint output into fitness or motion analysis pipelines
  • Serve as the pose-preprocessing step for virtual try-on or avatar workflows

Best For

  • App developers who want on-demand pose recognition inside DSH
  • Algorithm engineers working on motion or fitness analysis that relies on human keypoints
  • Teams building automation pipelines who need a unified pose API without maintaining their own inference stack