SiliconFlow-Vision-MCP
An MCP server that provides 8 vision tools for UI screenshot to code, OCR, error diagnosis, diagram understanding, data visualization analysis, UI diff, and image/video analysis, plus model list query, powered by SiliconFlow's multimodal API.
README
SiliconFlow-Vision-MCP
基于 SiliconFlow 多模态视觉 API(OpenAI 兼容格式)的 MCP 服务器,提供 8 个视觉工具 + 模型列表查询。
环境要求
- bun ≥ 1.0
- SiliconFlow API Key(https://cloud.siliconflow.cn/ 获取)
安装与启动
bun install # 安装依赖(已含 @modelcontextprotocol/sdk)
bun run src/index.js # 启动 MCP server(stdio)
配置 API Key
默认从环境变量 SILICONFLOW_API_KEY 读取。
Windows(PowerShell)
临时设置(仅当前会话):
$env:SILICONFLOW_API_KEY = "sk-你的Key"
永久设置(用户级环境变量):
setx SILICONFLOW_API_KEY "sk-你的Key"
Linux / macOS
临时设置:
export SILICONFLOW_API_KEY="sk-你的Key"
永久设置(写入 ~/.bashrc 或 ~/.zshrc):
echo 'export SILICONFLOW_API_KEY="sk-你的Key"' >> ~/.bashrc
source ~/.bashrc
自定义环境变量名
如需让 server 从其他变量读取 Key,设置 SILICONFLOW_API_KEY_NAME 指向目标变量:
- Windows:
$env:SILICONFLOW_API_KEY_NAME = "MY_KEY" - Linux/macOS:
export SILICONFLOW_API_KEY_NAME=MY_KEY
也可在 MCP 客户端配置中通过 env 注入(见下方示例)。
可选环境变量:
| 变量 | 默认值 | 说明 |
|---|---|---|
SILICONFLOW_API_KEY |
- | API Key(必填) |
SILICONFLOW_API_KEY_NAME |
SILICONFLOW_API_KEY |
指定从哪个环境变量读 Key |
SILICONFLOW_BASE_URL |
https://api.siliconflow.cn/v1 |
兼容服务地址 |
工具列表
| 工具 | 说明 |
|---|---|
ui_to_artifact |
UI 截图 → 代码 / 提示词 / 设计规范 / 描述 |
extract_text_from_screenshot |
截图 OCR 文字提取(代码、终端、文档) |
diagnose_error_screenshot |
错误弹窗/堆栈/日志截图 → 定位与修复建议 |
understand_technical_diagram |
架构图/流程图/UML/ER 图结构化解读 |
analyze_data_visualization |
仪表盘/图表 → 趋势、异常、业务要点 |
ui_diff_check |
两张 UI 截图对比,识别视觉差异 |
image_analysis |
通用图像理解(自由提问) |
video_analysis |
视频场景解析(本地 ≤8MB,MP4/MOV/M4V) |
list_models |
拉取远端可用多模态模型列表 |
所有视觉工具均支持 model 参数覆盖默认模型;默认模型 Qwen/Qwen3-VL-8B-Instruct(速度快、成本低;对质量要求高时可显式指定 Qwen/Qwen3-VL-32B-Instruct)。
内置模型能力(2026-08-04 对 SiliconFlow API 实测,发送最小图片/音频/视频探测,比文档更可靠;实测脚本见 scripts/probe-multimodal.mjs):
| 模型 | 图片 | 音频 | 视频 |
|---|---|---|---|
| Qwen/Qwen3-VL-8B-Instruct(默认) | ✅ | — | ✅ |
| Qwen/Qwen3-VL-32B-Instruct | ✅ | — | ✅ |
| Qwen/Qwen3-Omni-30B-A3B-Instruct | ✅ | ✅ | ✅ |
| zai-org/GLM-4.5V | ✅ | — | — |
| deepseek-ai/DeepSeek-OCR | ✅ | — | — |
| PaddlePaddle/PaddleOCR-VL-1.5 | ✅ | — | — |
注意:视频分析只能使用支持视频的模型(Qwen3-VL / Qwen3-Omni 系列);stepfun-ai/Step-3.5-Flash 经实测不是多模态模型,已从内置列表移除。
媒体输入支持三种形式:
- 本地文件路径(server 读取后转 base64,隐私:日志不记录路径与内容)
http(s)://URL(透传给远端)data:URI(base64)
大小限制:图片 20MB,视频 8MB(本地文件)。
注册到 Codex
在 ~/.codex/config.toml 中追加(将 <项目路径> 替换为实际路径):
[mcp_servers.siliconflow-vision-mcp]
command = "bun"
args = ["run", "<项目路径>/src/index.js"]
[mcp_servers.siliconflow-vision-mcp.env]
SILICONFLOW_API_KEY = "{env:SILICONFLOW_API_KEY}"
路径写法:
- Windows:
C:\Users\你的用户名\siliconflow-vision-mcp\src\index.js(TOML 中用单引号包裹,反斜杠无需转义) - Linux/macOS:
/home/你的用户名/siliconflow-vision-mcp/src/index.js
{env:...} 为 Codex 环境变量引用语法;也可以直接写入明文(不推荐)。
注册到其他 MCP 客户端(示例)
opencode ~/.config/opencode/opencode.jsonc:
{
"mcp": {
"siliconflow-vision-mcp": {
"type": "local",
"command": ["bun", "run", "<项目路径>/src/index.js"],
"enabled": true,
"environment": {
"SILICONFLOW_API_KEY": "{env:SILICONFLOW_API_KEY}"
}
}
}
}
验证
# 协议与工具注册检查(无需 API Key)
bun run scripts/verify.mjs
# 真实 API 调用检查(需要 SILICONFLOW_API_KEY)
bun run scripts/verify-live.mjs
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
Neon Database
MCP server for interacting with Neon Management API and databases
E2B
Using MCP to run code via e2b.
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.