# GLM-4.5V Vision

> Use this tool when you need to analyze and recognize images, extract metadata, or generate descriptions from visual content. It solves problems such as automated image processing, code documentation from screenshots, and AI-powered visual content analysis, accepting image URLs or local file paths as input and producing structured analysis results with confidence scoring as output. Ideal for developers seeking to integrate visual content analysis into their development tools and workflows.

Canonical page: https://skillsregistry.net/skills/lengbone-glm-vision  
JSON: https://api.skillsregistry.net/v1/skills/lengbone-glm-vision

## Description

This MCP server provides image recognition and analysis capabilities using the GLM-4.5V vision model, built with TypeScript and the Model Context Protocol SDK for integration with Claude and other AI assistants. The implementation offers specialized tools for analyzing various image types including general image description, code screenshot analysis with language detection and architecture insights, and automatic image processing with Sharp for format conversion and metadata extraction. Features include structured analysis results with confidence scoring, comprehensive error handling, and support for both local file paths and image URLs, making it valuable for developers who need AI-powered visual content analysis, code documentation from screenshots, or automated image processing workflows integrated into their development tools.

## Trust

- **Trust score (0–1):** 0.91
- **Verification tier:** verified
- **Last scanned:** 2026-09-28

## Facts

- **Version:** 1.0.0
- **Skill type:** atomic
- **Execution layer:** mcp-remote
- **Runtime environment:** api
- **Category:** media
- **Updated:** 2026-09-28

## Source

- **Source listing:** [PulseMCP](https://www.pulsemcp.com/servers/lengbone-glm-vision)
- **Repository:** <https://github.com/lengbone/mcp-vl>

## Use it

Resolve this record through the SkillsRegistry MCP server (no auth, read-only):

```
claude mcp add --transport http --scope user skillsregistry https://api.skillsregistry.net/mcp
```

```json
{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "get_skill",
    "arguments": {
      "slug": "lengbone-glm-vision"
    }
  }
}
```

REST: `GET https://api.skillsregistry.net/v1/skills/lengbone-glm-vision` · pull for local use: `GET https://api.skillsregistry.net/v1/skills/lengbone-glm-vision/pull`

---
SkillsRegistry indexes agent skills from public registries and GitHub. Skills we have analysed are scanned with Circle-IR and scored on six dimensions; each listing states its scan coverage. More: https://skillsregistry.net/llms.txt