Skip to main content
Glama
Sign In
GLM-4.5V MCP Server

GLM-4.5V MCP Server

by ccw33
OverviewSchemaSourceRelated ServersDiscussionsAdmin
TypeScript
Hybrid

GLM-4.5V MCP Server

GLM-4.5V 多模态能力的 MCP 服务器，提供图像处理、视觉查询和文件处理功能。

功能

  • read_image: 读取本地/URL图片并返回 dataURL 与尺寸信息

  • vision_query: 调用 GLM-4.5V 对图片进行 OCR/问答/检测

  • process_file: 使用 GLM-4.5V 处理文件（上传并提取内容）

Related MCP server: Look At Pic MCP

安装

npm install

配置

复制环境变量文件：

cp .env.example .env

编辑 .env 文件，填入你的 GLM API Key：

GLM_API_KEY=your_api_key_here
GLM_BASE_URL=https://open.bigmodel.cn/api/paas/v4/chat/completions

构建

npm run build

运行

npm start

或使用启动脚本：

./start-mcp.sh

开发模式

npm run dev

测试功能

创建测试文件并验证功能：

node scripts/test-file-processing.js

工具说明

read_image

读取图片并返回 base64 编码的 dataURL。

参数：

  • path (必需): 图片路径或URL

  • maxSide (可选): 最大边长，用于缩放

vision_query

使用 GLM-4.5V 分析图片。

参数：

  • path (必需): 图片路径或URL

  • prompt (必需): 查询提示词

  • mode (可选): 查询模式 ("describe", "ocr", "qa", "detect")

  • returnJson (可选): 是否返回JSON格式结果

process_file

使用 GLM-4.5V 处理文件，支持多种格式的文件上传和内容提取。

支持的文件格式：

  • 文档：PDF、DOCX、DOC、XLS、XLSX、PPT、PPTX、CSV、TXT

  • 图片：PNG、JPG、JPEG

文件大小限制：

  • 图片文件：最大 5MB

  • 其他文件：最大 50MB

参数：

  • filePath (必需): 本地文件路径

  • extractPrompt (可选): 内容提取提示词，用于指导如何提取文件内容

返回结果：

{
  "ok": true,
  "fileId": "file-xxx",
  "content": "提取的文件内容...",
  "fileType": "PDF文档",
  "filename": "document.pdf",
  "metadata": {
    "uploadTime": 1234567890,
    "fileSize": 1024000,
    "processingTime": 5000
  }
}

使用示例

处理 PDF 文档

# 通过 MCP 调用
{
  "tool": "process_file",
  "arguments": {
    "filePath": "./documents/report.pdf",
    "extractPrompt": "请提取文档中的主要内容和关键信息"
  }
}

处理 Excel 表格

{
  "tool": "process_file",
  "arguments": {
    "filePath": "./data/sales.xlsx",
    "extractPrompt": "请分析表格数据并总结销售趋势"
  }
}

Available Tools

3 tools
process_fileC

使用 GLM-4.5V 处理文件（上传并提取内容）。支持 PDF、DOCX、DOC、XLS、XLSX、PPT、PPTX、PNG、JPG、JPEG、CSV 等格式

Inputs: filePath (required), extractPrompt.

read_imageC

读取本地/URL图片并返回 dataURL 与尺寸信息

Inputs: path (required), maxSide.

vision_queryC

调用 GLM-4.5V 对图片进行 OCR/问答/检测

Inputs: path (required), prompt (required), mode, returnJson.

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    Deploy Server
    One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

    TDQS

    C2.9/5.0

    Scored across 3 tools

    Disambiguation2/5

    There is significant overlap between tools, particularly 'process_file' and 'read_image'/'vision_query'. 'process_file' handles image files among others, while 'read_image' and 'vision_query' specifically target images, creating ambiguity about which tool to use for image-related tasks. The descriptions do not clearly delineate boundaries, such as whether 'process_file' extracts text from images or if that's reserved for 'vision_query'.

    Naming Consistency4/5

    Tool names follow a consistent snake_case pattern (e.g., 'process_file', 'read_image', 'vision_query'), which is readable and predictable. However, there is a minor deviation in verb style: 'process' and 'read' are action-oriented, while 'vision_query' uses a noun-verb combination, slightly reducing consistency.

    Tool Count3/5

    With only 3 tools, the server feels thin for a vision/processing domain, potentially limiting functionality. While it covers basic file processing and image tasks, the low count may indicate missing operations for a comprehensive GLM-4.5V integration, such as text analysis or batch processing, making it borderline appropriate.

    Completeness2/5

    The tool set has significant gaps for a GLM-4.5V server. It lacks core operations like text querying, model configuration, or error handling tools. There is no clear coverage for non-image file types beyond extraction in 'process_file', and missing update/delete operations for processed data could lead to agent failures in complex workflows.

    Maintenance

    ActivityInactive
    ResponsivenessOutdated data

    Resources

    Related MCP Connectors

    Related MCP Servers

    • GLM Vision Server

      danilofalcao
      A
      license
      B
      quality
      D
      maintenance
      Enables image analysis using GLM-4.5V's vision capabilities from Z.AI. Supports analyzing both local image files and URLs with customizable prompts and parameters.
      1
      7
      MIT

    • glm-vision-mcp

      kkdnn
      A
      license
      Not graded
      quality
      D
      maintenance
      Provides image understanding and OCR via GLM-4.6V-Flash, supporting URL, base64, and local file inputs. Enables AI assistants to analyze images and extract text from screenshots, documents, and more.
      108 npm
      MIT