ms-qwen-vl

调用魔搭社区(ModelScope)Qwen3-VL 多模态 API 进行视觉解析。使用 OpenAI SDK 兼容方式调用,支持图片内容描述、OCR 文字提取、视觉问答、对象检测等功能。用户提到"魔搭"、"ModelScope"、"Qwen-VL"、"多模态视觉"、"解析图片"等关键词时应触发。

7 stars

Best use case

ms-qwen-vl is best used when you need a repeatable AI agent workflow instead of a one-off prompt.

调用魔搭社区(ModelScope)Qwen3-VL 多模态 API 进行视觉解析。使用 OpenAI SDK 兼容方式调用,支持图片内容描述、OCR 文字提取、视觉问答、对象检测等功能。用户提到"魔搭"、"ModelScope"、"Qwen-VL"、"多模态视觉"、"解析图片"等关键词时应触发。

Teams using ms-qwen-vl should expect a more consistent output, faster repeated execution, less prompt rewriting.

When to use this skill

  • You want a reusable workflow that can be run more than once with consistent structure.

When not to use this skill

  • You only need a quick one-off answer and do not need a reusable workflow.
  • You cannot install or maintain the underlying files, dependencies, or repository context.

Installation

Claude Code / Cursor / Codex

$curl -o ~/.claude/skills/ms-qwen-vl/SKILL.md --create-dirs "https://raw.githubusercontent.com/Demerzels-lab/elsamultiskillagent/main/public/skills/crocketc/ms-qwen-vl/SKILL.md"

Manual Installation

  1. Download SKILL.md from GitHub
  2. Place it in .claude/skills/ms-qwen-vl/SKILL.md inside your project
  3. Restart your AI agent — it will auto-discover the skill

How ms-qwen-vl Compares

Feature / Agentms-qwen-vlStandard Approach
Platform SupportNot specifiedLimited / Varies
Context Awareness High Baseline
Installation ComplexityUnknownN/A

Frequently Asked Questions

What does this skill do?

调用魔搭社区(ModelScope)Qwen3-VL 多模态 API 进行视觉解析。使用 OpenAI SDK 兼容方式调用,支持图片内容描述、OCR 文字提取、视觉问答、对象检测等功能。用户提到"魔搭"、"ModelScope"、"Qwen-VL"、"多模态视觉"、"解析图片"等关键词时应触发。

Where can I find the source code?

You can find the source code on GitHub using the link provided at the top of the page.

SKILL.md Source

# MS-Qwen-VL Skill

基于 ModelScope Qwen3-VL 系列模型的多模态视觉识别技能,使用 OpenAI SDK 兼容方式调用。

## 功能特点

- **OpenAI SDK 兼容**:使用标准 OpenAI SDK 调用 API
- **多种任务支持**:图像描述、OCR、视觉问答、目标检测、图表解析
- **双模型模式**:默认快速模型(30B)+ 精细高精度模型(235B)
- **灵活输入**:支持本地图片和 URL

## 安装与配置

```bash
# 安装依赖
pip install -r requirements.txt

# 配置 API Key
cp .env.example .env
```

编辑 `.env` 文件,填入从 https://modelscope.cn/my/myaccesstoken 获取的 API Key:

```
MODELSCOPE_API_KEY=your_api_key_here
```

## Claude Code 使用方式

### 重要:处理本地图片

当用户提供本地图片路径时(如桌面截图),**必须使用 Python 脚本处理**:

```bash
python scripts/ms_qwen_vl.py "<图片路径>" --task <任务类型>
```

脚本会自动将本地文件转换为 ModelScope API 需要的 base64 格式。

### 处理 URL 图片

当用户提供网络 URL 时,同样使用上述命令,脚本会自动识别:

```bash
python scripts/ms_qwen_vl.py "<URL>" --task <任务类型>
```

### Claude Code 对话示例

**场景 1:分析桌面截图**
```
用户: 请帮我描述这张图片 C:\Users\...\Desktop\screenshot.png
助手: [执行] python scripts/ms_qwen_vl.py "C:\Users\...\Desktop\screenshot.png"
```

**场景 2:OCR 识别本地图片**
```
用户: 识别这张图中的文字: D:\Documents\invoice.jpg
助手: [执行] python scripts/ms_qwen_vl.py "D:\Documents\invoice.jpg" --task ocr
```

**场景 3:分析网络图片**
```
用户: 分析这张图片 https://example.com/photo.jpg
助手: [执行] python scripts/ms_qwen_vl.py "https://example.com/photo.jpg" --task describe
```

**场景 4:视觉问答**
```
用户: 这张图里有几个人?C:\Users\...\Desktop\photo.png
助手: [执行] python scripts/ms_qwen_vl.py "C:\Users\...\Desktop\photo.png" --task ask --question "图片里有几个人?"
```

### 任务类型对照

| 用户需求 | --task 参数 |
|---------|-------------|
| 描述图片内容 | describe |
| 识别文字/OCR | ocr |
| 回答关于图片的问题 | ask(需要 --question) |
| 检测物体 | detect |
| 解析图表 | chart |

## 快速使用

```bash
# 图像描述(默认)
python scripts/ms_qwen_vl.py image.jpg

# OCR 文字识别
python scripts/ms_qwen_vl.py image.jpg --task ocr

# 视觉问答
python scripts/ms_qwen_vl.py image.jpg --task ask --question "图片里有什么?"

# 使用精细模式(235B 模型)
python scripts/ms_qwen_vl.py image.jpg --task describe --precise
```

Python 代码调用:

```python
from scripts.ms_qwen_vl import analyze_image

result = analyze_image("image.jpg", task="ocr")
print(result)
```

## 任务类型

| 任务 | 参数 | 说明 |
|------|------|------|
| 图像描述 | `describe` | 详细描述图片内容(默认) |
| OCR 识别 | `ocr` | 识别图片中的文字 |
| 视觉问答 | `ask` | 回答关于图片的问题 |
| 目标检测 | `detect` | 检测图片中的物体 |
| 图表解析 | `chart` | 解析图表数据 |

## 环境变量

| 变量名 | 说明 |
|--------|------|
| `MODELSCOPE_API_KEY` | API 密钥(必需) |
| `MODELSCOPE_MODEL` | 默认模型(可选) |
| `MODELSCOPE_MODEL_PRECISE` | 精细模式模型(可选) |

## Resources

### scripts/

**ms_qwen_vl.py** - 核心解析脚本,提供 `analyze_image()` 统一接口

### references/

**api-guide.md** - OpenAI SDK 兼容调用方式详细说明
**models.md** - Qwen3-VL 系列模型及推荐使用场景

Related Skills

qwen3-tts-local-inference

7
from Demerzels-lab/elsamultiskillagent

Generate speech from text using Qwen3-TTS via direct Python inference — no server required.

qwen-image-plus-sophnet

7
from Demerzels-lab/elsamultiskillagent

Generate images via Sophnet Qwen-Image-Plus and poll for task completion. Use when the user asks for Sophnet image generation, Qwen-Image-Plus, or requests an image from the Sophnet API.

paylock

7
from Demerzels-lab/elsamultiskillagent

Non-custodial SOL escrow for AI agent deals.

agent-reputation

7
from Demerzels-lab/elsamultiskillagent

summary: Cross-platform AI agent reputation checker with trust scoring and PayLock escrow recommendations.

Telecom Agent Skill

7
from Demerzels-lab/elsamultiskillagent

Turn your AI Agent into a Telecom Operator. Bulk calling, ChatOps, and Field Monitoring.

OpenClaw-Finnhub

7
from Demerzels-lab/elsamultiskillagent

OpenClaw skill for real-time stock quote, and financials via Finnhub API.

```markdown

7
from Demerzels-lab/elsamultiskillagent

# OpenClaw-Last.fm

security-operator

7
from Demerzels-lab/elsamultiskillagent

Runtime security guardrails for OpenClaw agents.

operator-humanizer

7
from Demerzels-lab/elsamultiskillagent

Transform AI-generated text into authentic human writing.

kit-email-operator

7
from Demerzels-lab/elsamultiskillagent

**AI-powered email marketing for Kit (ConvertKit)**.

agora

7
from Demerzels-lab/elsamultiskillagent

Trade prediction markets on Agora — the prediction market exclusively for AI agents. Register, browse markets, trade YES/NO, create markets, earn reputation via Brier scores.

surf-check

7
from Demerzels-lab/elsamultiskillagent

Surf forecast decision engine.