{"_id":"@cqsjjb/mcp-vision-understanding","name":"@cqsjjb/mcp-vision-understanding","dist-tags":{"latest":"1.0.0"},"versions":{"1.0.0":{"name":"@cqsjjb/mcp-vision-understanding","version":"1.0.0","description":"MCP server for vision understanding - 支持图片上传和分析功能","type":"module","main":"index.js","bin":{"mcp-vision-understanding":"index.js"},"scripts":{"start":"node index.js","test":"node test.js"},"keywords":["mcp","vision","vision-understanding","image-analysis","model-context-protocol","visual-model"],"author":"","license":"MIT","repository":{"type":"git","url":""},"engines":{"node":">=18.0.0"},"dependencies":{"@modelcontextprotocol/sdk":"^0.5.0","axios":"^1.12.2"},"_id":"@cqsjjb/mcp-vision-understanding@1.0.0","gitHead":"2dfc96d1a01f2aa19476d2741150bfc78b3b93ae","_nodeVersion":"22.12.0","_npmVersion":"10.9.0","dist":{"integrity":"sha512-xHEpRTt8Zmh+1YSN3nsjJEz2jUIWeWN96r/XYJqZdac4V+jYAotN4RraTy7HdVbeLLfmBRFRyAbJlC7rYtls0A==","shasum":"00285ae8f1961eeb243d4f5e959fabdb260af0bf","tarball":"https://registry.npmjs.org/@cqsjjb/mcp-vision-understanding/-/mcp-vision-understanding-1.0.0.tgz","fileCount":4,"unpackedSize":16277,"signatures":[{"keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U","sig":"MEUCIHkYAV2gwZkLzgBR5H39AeCBvXgM1ZSMalfbpYKx8FQHAiEAvSDmw1zbKNQNVT+5WItrJ6kth4tsqEbpxaem3nRy+KQ="}]},"_npmUser":{"name":"cqjjb","email":"995934055@qq.com"},"directories":{},"maintainers":[{"name":"cqjjb","email":"995934055@qq.com"}],"_npmOperationalInternal":{"host":"s3://npm-registry-packages-npm-production","tmp":"tmp/mcp-vision-understanding_1.0.0_1764571858826_0.5328165120438846"},"_hasShrinkwrap":false}},"time":{"created":"2025-12-01T06:50:58.730Z","1.0.0":"2025-12-01T06:50:59.008Z","modified":"2025-12-01T06:50:59.341Z"},"maintainers":[{"name":"cqjjb","email":"995934055@qq.com"}],"description":"MCP server for vision understanding - 支持图片上传和分析功能","keywords":["mcp","vision","vision-understanding","image-analysis","model-context-protocol","visual-model"],"repository":{"type":"git","url":""},"license":"MIT","readme":"# Vision Understanding MCP Server\r\n\r\n这是一个视觉理解 MCP (Model Context Protocol) 服务器，用于在 Cursor 等 IDE 中分析图片内容。支持多种视觉理解模型。\r\n\r\n## 功能特性\r\n\r\n- ✅ 支持图片上传和分析\r\n- ✅ 支持自定义提示词\r\n- ✅ 支持多种图片格式（jpg, jpeg, png, gif, webp, bmp）\r\n- ✅ 详细的错误处理和反馈\r\n\r\n## 安装\r\n\r\n### 方式一：使用 NPM 包（推荐，全局可用）\r\n\r\n发布到 NPM 后，可以直接使用：\r\n\r\n```bash\r\n# 全局安装（可选）\r\nnpm install -g mcp-vision-understanding\r\n```\r\n\r\n或者在 Cursor 配置中使用 `npx`（无需安装）：\r\n\r\n```json\r\n{\r\n  \"mcpServers\": {\r\n    \"vision-understanding\": {\r\n      \"command\": \"npx\",\r\n      \"args\": [\"-y\", \"mcp-vision-understanding\"],\r\n      \"env\": {\r\n        \"VISION_API_KEY\": \"your-api-key-here\",\r\n        \"VISION_MODEL\": \"your-model-name\"\r\n      }\r\n    }\r\n  }\r\n}\r\n```\r\n\r\n### 方式二：本地安装\r\n\r\n```bash\r\ncd mcp-vision-understanding\r\nnpm install\r\n```\r\n\r\n## 配置\r\n\r\n### 环境变量配置（推荐）\r\n\r\n使用环境变量配置 API 密钥，更安全：\r\n\r\n- `VISION_API_KEY`: API 密钥（必需）\r\n- `VISION_BASE_URL`: API 基础 URL（可选，默认：https://dashscope.aliyuncs.com/compatible-mode/v1）\r\n- `VISION_MODEL`: 模型名称（必需）\r\n\r\n### 在 Cursor 中配置\r\n\r\n**使用 NPM 包（推荐）**：\r\n\r\n```json\r\n{\r\n  \"mcpServers\": {\r\n    \"vision-understanding\": {\r\n      \"command\": \"npx\",\r\n      \"args\": [\"-y\", \"mcp-vision-understanding\"],\r\n      \"env\": {\r\n        \"VISION_API_KEY\": \"your-api-key-here\",\r\n        \"VISION_MODEL\": \"your-model-name\"\r\n      }\r\n    }\r\n  }\r\n}\r\n```\r\n\r\n**使用本地安装**：\r\n\r\n```json\r\n{\r\n  \"mcpServers\": {\r\n    \"vision-understanding\": {\r\n      \"command\": \"node\",\r\n      \"args\": [\"E:/path/to/mcp-vision-understanding/index.js\"],\r\n      \"env\": {\r\n        \"VISION_API_KEY\": \"your-api-key-here\",\r\n        \"VISION_MODEL\": \"your-model-name\"\r\n      }\r\n    }\r\n  }\r\n}\r\n```\r\n\r\n**注意**: \r\n- 使用 NPM 包方式，所有项目都可以使用，无需每个项目都安装\r\n- 使用本地安装方式，需要将路径替换为您项目的实际路径\r\n\r\n## 使用方法\r\n\r\n在 Cursor 中，您可以通过以下方式使用：\r\n\r\n1. **上传图片并分析**:\r\n   ```\r\n   使用 analyze_image 工具分析图片: 无标题.png\r\n   ```\r\n\r\n2. **使用自定义提示词**:\r\n   ```\r\n   分析图片 无标题.png，请告诉我图片中有哪些物体和它们的颜色\r\n   ```\r\n\r\n## 工具说明\r\n\r\n### analyze_image\r\n\r\n分析图片内容并返回详细描述。\r\n\r\n**参数**:\r\n- `imagePath` (必需): 图片文件的路径\r\n- `prompt` (可选): 自定义提示词，默认为\"请详细描述这张图片的内容\"\r\n\r\n**示例**:\r\n```json\r\n{\r\n  \"imagePath\": \"无标题.png\",\r\n  \"prompt\": \"请详细描述这张图片的内容，包括主要物体、颜色、场景等\"\r\n}\r\n```\r\n\r\n## 支持的图片格式\r\n\r\n- JPEG (.jpg, .jpeg)\r\n- PNG (.png)\r\n- GIF (.gif)\r\n- WebP (.webp)\r\n- BMP (.bmp)\r\n\r\n## 注意事项\r\n\r\n1. 确保您的 API 密钥有效且有足够的配额\r\n2. 图片文件路径可以是相对路径或绝对路径\r\n3. 相对路径会从当前工作目录解析\r\n4. API 调用超时时间为 60 秒\r\n\r\n## 故障排除\r\n\r\n### 图片文件不存在\r\n- 检查文件路径是否正确\r\n- 确保使用正确的路径分隔符（Windows 使用 `\\`，Unix 使用 `/`）\r\n\r\n### API 调用失败\r\n- 检查 API 密钥是否正确\r\n- 检查网络连接\r\n- 查看控制台错误信息\r\n\r\n### 不支持的文件格式\r\n- 确保图片文件扩展名在支持列表中\r\n- 可以尝试将图片转换为支持的格式\r\n\r\n## 许可证\r\n\r\nMIT\r\n\r\n","readmeFilename":"README.md","_rev":"1-2c3737f85fb67930dca4f8f520d75651"}