Snapgrab

Snapgrab

한국어 문서 · llms.txt

将 URL 转换为带元数据的截图。针对 Claude Vision 进行了优化。

from snapgrab import capture

result = await capture("https://example.com")
print(result.path)           # /tmp/snapgrab/example_com_desktop_20260317_120000.png
print(result.metadata.title) # "Example Domain"
print(result.vision_tokens)  # ~2764

功能

• 截图 — 支持整页截图，可保存为 PNG、JPEG、PDF
• 页面元数据 — 标题、描述、Open Graph 标签、网站图标、HTTP 状态
• 针对 Claude Vision 优化 — 自动调整为 1568px，并估算令牌成本
• 视口预设 — 桌面端（1920x1080）、平板（768x1024）、移动端（375x812）
• 内置 MCP 服务 — 为 Claude Code / MCP 客户端提供 3 个工具
• 元素截图 — 截取指定 CSS 选择器对应的内容
• 深色模式 — 强制使用浅色或深色配色方案

安装

pip install snapgrab

首次运行时会提示安装 Playwright 浏览器：

playwright install chromium

用法

Python API

import asyncio
from snapgrab import capture

async def main():
    # Basic screenshot
    result = await capture("https://example.com")

    # Mobile viewport, full page
    result = await capture("https://example.com", viewport="mobile", full_page=True)

    # Dark mode, JPEG format
    result = await capture("https://example.com", dark_mode=True, format="jpeg")

    # Specific element
    result = await capture("https://example.com", selector="#main-content")

    # Custom viewport
    result = await capture("https://example.com", viewport=(1440, 900))

asyncio.run(main())

CLI

snapgrab https://example.com                          # basic PNG
snapgrab https://example.com -v mobile -f             # mobile, full page
snapgrab https://example.com --format jpeg -q 90      # JPEG quality 90
snapgrab https://example.com -s "#hero" --dark-mode   # element + dark mode
snapgrab https://example.com -j                       # JSON output
snapgrab meta https://example.com                     # metadata only

MCP 服务

pip install "snapgrab[mcp]"
snapgrab-mcp  # starts stdio MCP server

工具：

• capture_screenshot — 截取 URL 并获取元数据和视觉令牌估算
• capture_comparison — 比较桌面端与移动端（或任意视口）
• extract_page_metadata — 仅提取元数据，不截取屏幕

工作原理

flowchart LR
    A["🔗 URL"] --> B["Playwright\nlaunch browser"]
    B --> C["📸 Screenshot\nfull page / viewport"]
    C --> D["Resize for\nClaude Vision"]
    D --> E["✅ Image +\nMetadata"]

CaptureResult

result.path            # saved file path
result.format          # "png", "jpeg", "pdf"
result.width           # viewport width
result.height          # page height (full_page) or viewport height
result.file_size       # bytes
result.vision_tokens   # estimated Claude Vision token cost
result.vision_path     # path to Vision-optimized image (≤1568px)
result.processing_time_ms
result.metadata.title
result.metadata.description
result.metadata.og_title
result.metadata.og_image
result.metadata.favicon_url
result.metadata.status_code
result.metadata.url    # final URL after redirects

使用项目

• watchdeck — 通过视觉差异和安全防护监控网页

许可证

MIT

---

_{属于 QuartzUnit 生态系统 — 一组可组合的 Python 库，用于数据收集、提取、搜索和 AI Agent 安全。}