glmv-grounding

Extract and visualize GLM-V grounding results from images or videos.

Updated Apr 2, 2026
One-click install
npx skills add https://github.com/ViewWay/openclaw-skills --skill glmv-grounding-viewway
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: glmv-grounding
Source: https://github.com/ViewWay/openclaw-skills/tree/main/glmv-grounding
Command: npx skills add https://github.com/ViewWay/openclaw-skills --skill glmv-grounding-viewway

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires requests, Pillow, opencv-python, numpy, matplotlib, decord, and includes scripts (resource) components.

What problem does it solve?

GLMV-Grounding provides automated grounding results for prompts describing targets in images or videos, and visualizes coordinates and trackers to assist analysis.

Core Features & Use Cases

  • Ground targets in images using 2D bounding boxes, 2D points, and 3D boxes, with relative coordinates (0-1000) and optional tracking in video.
  • Visualize grounding results on images or videos and export visualizations for reporting.
  • Use cases: model evaluation, QA checks, development of AI agents that reason about scene content.

Quick Start

Run the GLM-V grounding CLI with an image or video and prompt to obtain grounding results and optional visualizations.

Frequently Asked Questions about glmv-grounding

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I extract and visualize 3D bounding boxes from images using GLM-V?

To extract and visualize 3D bounding boxes from images using GLM-V, the Skill processes your prompt and image to return relative coordinates alongside exported visualizations for scene analysis.

What is video tracking with coordinate conversion for computer vision tasks?

Video tracking with coordinate conversion is the process of identifying targets across frames and mapping relative coordinates to visualize trackers, which this Skill automates using GLM-V for video surveillance analysis.

Do I need a ZHIPU_API_KEY to use GLM-V for grounding targets in videos?

Yes, you need a ZHIPU_API_KEY to access the fixed GLM-V API endpoint for grounding targets in videos, enabling extraction of structured grounding results and visual trackers.

Can I use local files and public URLs for image processing with GLM-V grounding?

Yes, you can use both local files and public URLs for image processing with GLM-V grounding, allowing you to process diverse inputs and generate visualizations for model evaluation or QA checks.

What are the limitations of using GLM-V for 2D point grounding in image analysis?

A limitation of using GLM-V for 2D point grounding is that relative coordinates are constrained to a 0-1000 scale, requiring coordinate conversion for precise pixel mapping in detailed image analysis tasks.