polydoc-mcp
MCP server that converts HTML or URLs to PDF, captures screenshots, and generates EU-compliant e-invoices (Factur-X/ZUGFeRD).
README
polydoc-mcp
An MCP server for PolyDoc, a REST API that converts HTML or URLs to PDF, captures screenshots, and generates EU-compliant e-invoices (Factur-X / ZUGFeRD hybrid PDF/A-3). It lets MCP clients (Claude Desktop, Claude Code, Cursor, and others) drive PolyDoc directly.
Tools
polydoc_html_to_pdf- HTML, URL, or saved template to PDF (layout, margins, page format, page ranges, bookmarks, accessible/tagged PDFs).polydoc_screenshot- HTML, URL, or template to PNG / JPEG / WebP, with viewport and device-pixel-ratio control. Returns an inline image preview when small enough.polydoc_generate_einvoice- Factur-X or ZUGFeRD hybrid PDF/A-3 from structured invoice data, profiles fromminimumtoextended.polydoc_test_credentials- verify the configured API key with a minimal sandbox render (never draws production quota).
Content can come from a URL, an inline HTML string, or a saved template (with Liquid templateData). Downloads are written to a local directory and the file path is returned; you can also deliver to your cloud storage (presigned URL) or a webhook.
Installation
Requires Node.js >= 20.15. Published to npm, so clients can run it with npx:
npx -y polydoc-mcp
The server is configured entirely through environment variables (see below).
Configuration
| Variable | Required | Default | Description |
|---|---|---|---|
POLYDOC_API_KEY |
yes | - | API key from dashboard.polydoc.tech. |
POLYDOC_SANDBOX |
no | false |
Default sandbox mode (watermarked, rate-limited). A tool call can override per request. |
POLYDOC_BASE_URL |
no | https://api.polydoc.tech |
Override for self-hosted or staging. |
POLYDOC_OUTPUT_DIR |
no | OS temp dir + /polydoc |
Where downloads are written. Output never escapes this folder. |
Client setup
Claude Desktop
Add to claude_desktop_config.json (Settings > Developer > Edit Config):
{
"mcpServers": {
"polydoc": {
"command": "npx",
"args": ["-y", "polydoc-mcp"],
"env": {
"POLYDOC_API_KEY": "your-key",
"POLYDOC_OUTPUT_DIR": "/Users/you/Documents/polydoc"
}
}
}
}
Claude Code
claude mcp add polydoc -e POLYDOC_API_KEY=your-key -- npx -y polydoc-mcp
Cursor
Add to ~/.cursor/mcp.json (or a project .cursor/mcp.json) using the same mcpServers block shown for Claude Desktop.
Output and delivery
- Download (default): the file is written under
POLYDOC_OUTPUT_DIRand the absolute path is returned with metadata (size, conversion id, credits). Screenshots also return an inline image block. PassreturnBase64: trueto also receive the bytes as base64 (heavy; for clients without a filesystem). - Cloud storage: set
delivery: "cloudStorage"andpresignedUrl. The upload URL is returned. - Webhook: set
delivery: "webhook"and awebhookobject ({ url, async?, method?, headers? }).
Anything not in a tool's schema?
Every conversion tool has an advanced object that is deep-merged into the request body, so any API capability not surfaced as a typed field (for example pdf.watermark, pdf.encryption, render, request) is still reachable. See the field reference at docs.polydoc.tech.
Examples
Worked tool inputs for each angle live in examples/: a branded PDF from a template, a URL screenshot, and a ZUGFeRD / EN 16931 e-invoice.
Development
yarn installyarn build- compile todist/yarn lintyarn test- unit tests (request builder, output jail, tool handlers) and an in-memory MCP round-tripPOLYDOC_API_KEY=your-key yarn test:integration- live smoke tests against the sandboxyarn scrub:check- check for em-dashes in source and docs
Run the built server directly for a quick check:
POLYDOC_API_KEY=your-key node dist/index.js
License
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。