Snip
Snip
Claude Code 的视觉模式。 也适用于 Cursor、Windsurf、Cline 以及任何 AI 编码 CLI。
安装 Snip,运行 snip setup,你的 AI 编码智能体就会开始渲染图表和预览,而不是用文本描述它们。架构?序列图。数据模型?ER 图。UI 组件?渲染预览。一切皆为视觉化,一切皆为自动化。
你可以审查 Claude 渲染的内容——批准它,或者直接在屏幕上进行标注。圈出错误的地方,添加箭头,写下备注。Claude 会看到你的标注并进行迭代。
工作原理
Snip 作为菜单栏/系统托盘应用程序运行,并带有一个可供任何 AI 编码智能体直接调用的 CLI:
# Claude writes a Mermaid diagram to a file, then renders it for your review
snip render --format mermaid < architecture.mmd
# Claude renders an HTML component preview
snip render --format html < preview.html
# Open any image for annotated review
snip open screenshot.png --message "Is the layout correct?"
# Search your screenshot library
snip search "login page error"智能体会收到结构化的 JSON 返回值:{ status: "approved" | "changes_requested", edited, path, text }。你可以进行空间标注、输入文本反馈,或者直接批准。
专为 Claude Code 构建。也适用于 Cursor、Windsurf、Cline 以及任何可以运行 shell 命令的工具。为没有 shell 访问权限的智能体提供 MCP 服务器。
Related MCP server: Screenshot MCP Server
安装
macOS (Homebrew)
brew install --cask rixinhahaha/snip/snip或者从 Releases 下载 DMG (Apple Silicon)。
Linux
从 Releases 下载:
AppImage (便携式,适用于任何发行版) —
Snip-x.y.z-x86_64.AppImagedeb (Ubuntu/Debian) —
Snip-x.y.z-amd64.deb
与 Claude Code 一起使用
snip setup就是这样。snip setup 会自动配置 Claude Code 以使用视觉输出——包括规则、/diagram 技能和权限。下次你向 Claude 询问架构、数据模型或流程时,它会渲染图表而不是用文本描述。
在任何 Claude Code 对话中尝试 /diagram 来可视化你正在讨论的内容。
CLI 参考
命令 | 功能 |
| 为 Claude Code (以及 Cursor、Windsurf、Cline) 启用视觉模式 |
| 从 stdin 渲染 Mermaid 图表,并打开以供审查 |
| 从 stdin 渲染 HTML,并打开以供审查 |
| 打开任何图像进行标注审查 |
| 按描述搜索截图库 |
| 通过 OCR 从图像中提取文本 |
| 列出所有带有元数据的已保存截图 |
| 获取特定截图的元数据 |
| 将截图加入 AI 分类队列 |
| 列出所有类别 |
所有审查命令 (render, open) 都会阻塞直到用户完成,并返回结构化的 JSON。
MCP 服务器
对于没有 shell 访问权限的智能体 (Claude Desktop, 托管环境),Snip 也提供了一个 MCP 服务器:
{
"mcpServers": {
"snip": {
"command": "node",
"args": ["/path/to/snip/src/mcp/server.js"]
}
}
}MCP 服务器公开了相同的功能:render_diagram, open_in_snip, search_screenshots, list_screenshots, get_screenshot, transcribe_screenshot, organize_screenshot, get_categories, install_extension。
同时也是一款截图工具
Snip 本身就是一个完整的截图 + 标注应用程序:
Cmd+Shift+2 (macOS) / Ctrl+Shift+2 (Linux) — 通过区域选择或窗口捕捉进行捕获
Cmd+Shift+1 / Ctrl+Shift+1 — 快速截图 (直接捕获到剪贴板)
标注 — 矩形、箭头、文本、标签、模糊画笔、AI 分割
Esc — 复制到剪贴板并关闭
Cmd+S — 保存到磁盘 + AI 在后台进行整理
AI 整理使用本地视觉 LLM (通过 Ollama) 来命名、分类和标记每一张保存的截图。语义搜索可以通过描述截图内容来查找任何截图。
快捷键
在 Linux 上,将 Cmd 替换为 Ctrl。
快捷键 | 操作 |
Cmd+Shift+2 | 捕获截图 |
Cmd+Shift+1 | 快速截图 (选择并复制到剪贴板) |
Cmd+Shift+S | 打开语义搜索 |
Cmd+S | 保存到磁盘 (在编辑器中) |
Esc / Enter | 复制到剪贴板并关闭 (在编辑器中) |
V / C / R / T / A / G / B / S | 选择 / 裁剪 / 矩形 / 文本 / 箭头 / 标签 / 模糊 / 分割工具 |
Cmd+Z | 撤销 (先撤销标注,然后是裁剪,最后是分割剪切) |
开发
npm install
npm run rebuild # compile native modules (macOS)
npm start # launch (tray icon appears)需要 macOS 14+ 或 Linux (Wayland),Node.js 18+。
文档
文档 | 内容 |
愿景、功能规格、术语 | |
调色板、组件模式、玻璃效果 | |
代码结构、IPC 通道、数据流 | |
构建流水线、签名、原生模块 | |
分步流程、边缘情况、测试用例 |
技术栈
Electron 33 / Fabric.js 7 / Mermaid.js 11 / Ollama (本地 LLM) / HuggingFace Transformers.js / SlimSAM (ONNX)
所有 AI 均在本地运行——核心功能无需云端 API。
赞助
Snip 是免费、开源的,由两名开发者构建。如果它对你的工作流程有帮助,请考虑赞助该项目。
赞助有助于支付:
Apple 开发者计划 — 代码签名和公证所必需,以防止 macOS 阻止该应用程序
CI/CD 成本 — macOS 和 Linux 的自动化构建、测试和发布流水线
持续开发 — 新的标注工具、更好的 AI 集成、Linux 改进和错误修复
每一份贡献都有助于保持 Snip 的维护和改进。谢谢!
许可证
MIT
This server cannot be deployed
Maintenance
Related MCP Connectors
Comment on AI-generated webpages; feedback flows back to your coding agent. Free, MIT, local-first.
SnapRender gives AI agents eyes on the web through a hosted MCP server. Send a public URL and get the page back as an image, a PDF or clean markdown, with no browser to run or maintain. Capture the viewport or the full page as PNG, JPEG, WebP or PDF, at desktop size or with a phone, tablet or laptop preset, in light or dark mode. Ads and cookie banners are removed by default, and you can hide elements or click one before the capture. extract_content reads the same page as markdown, text, article, links or metadata, so an agent can see a page and read it in one session. Made for monitoring and client work: batch up to 50 URLs in one job, check the cache for free before spending a capture, share signed links that render without an API key, and get webhooks when a batch finishes or a scheduled capture spots a change. Track competitors' pricing and landing pages, or send clients current screenshots of their sites. Connect with OAuth (Google, GitHub or email) or an API key. Works with Claude, ChatGPT, Cursor, VS Code and any MCP client. Free plan: 200 screenshots a month, no credit card. Paid plans from $9 a month for 2,000 up to $199 a month for 200,000. Home: https://snap-render.com Docs: https://snap-render.com/ai-agents
Grabbit gives AI agents eyes on the web through a hosted MCP server. Send a public URL and get a pixel-perfect hosted image back, without maintaining Chromium, Playwright, or a browser fleet. Capture a full page, exact viewport, or single CSS selector as PNG, JPEG, or WebP. Grabbit handles cookie and consent banners, waits for JavaScript-heavy pages, blocks private and internal URLs, supports safe retries with idempotency keys, and delivers async results through signed webhooks. Completed captures include a CDN URL. Connect with OAuth 2.1 or an API key. Grabbit works with Claude, Cursor, Codex, and any MCP client. Live captures cost $0.002 each. The $50 annual plan includes 25,000 prepaid credits that never reset or expire. Free test keys return placeholder images, so you can wire up the integration before paying. Home: https://grabbit.live Docs: https://grabbit.live/screenshot-api Built by BrainGrid.
Turn calls, notes, and code into process flow diagrams. Send them to Jira, Notion, and more.
1
Related MCP Servers
- AlicenseAqualityAmaintenanceA macOS utility that captures screenshots and analyzes them with AI vision, enabling AI assistants to see and interpret what's on your screen.32,586 npm5,294MIT
- FlicenseNot gradedqualityDmaintenanceA Model Context Protocol server that enables natural language-driven screenshot capture on macOS, allowing users to take full desktop screenshots, capture specific windows, or select custom screen areas through Claude Desktop.-
- AlicenseAqualityDmaintenanceEnables AI agents to capture and analyze screenshots of macOS applications, windows, or the entire screen using local (Ollama) or cloud-based AI vision models, with non-intrusive, fast screen capture via Apple's ScreenCaptureKit.312 npm2MIT
- AlicenseAqualityDmaintenanceMCP server for FreezeText — OCR anything on your Mac screen from Claude, Cursor, or any MCP client. Freeze the screen and extract text via Apple Vision (videos, popups, protected PDFs), OCR a region or a base64 image, and manage a searchable capture history. 12 tools. Bridge open-source (MIT), FreezeText app is free.121MIT