nudocs-mcp-server

nudocs-mcp-server

Enables document processing and editing using Nudocs.ai, including listing, reading, uploading, downloading, and modifying documents through natural language.

Category
访问服务器

README

Nudocs MCP Server

An MCP (Model Context Protocol) server for Nudocs.ai - a document processing service by Nutrient.

Features

This MCP server provides the following tools for working with Nudocs documents:

  • list_documents - List all Nudocs documents
  • read_document - Read document content as Markdown
  • get_document_access_link - Get a shareable access link to a document
  • download_document - Download documents in various formats
  • upload_file - Upload files to Nudocs
  • delete_document - Delete documents by ID
  • list_top_level_elements - List top-level elements with IDs and text previews
  • find_elements_by_text - Find element IDs by visible text
  • replace_paragraph_text - Replace paragraph text while preserving inline formatting when possible
  • insert_paragraph_text - Insert one plain-text paragraph at an anchor
  • insert_table - Insert a table from plain-text cells
  • move_element - Move one paragraph/table before or after a target paragraph
  • delete_element - Delete one paragraph, table, or image by ID
  • set_paragraph_style - Set paragraph style (normal, headings, bullet/number list)
  • insert_image - Insert one image from URL at an anchor

Recommended Edit Workflow

  1. Discover IDs with list_top_level_elements or find_elements_by_text.
  2. Apply the specific edit tool for the operation you want.
  3. Verify result with read_document.

All editing tools use object-only payloads and simple, LLM-friendly inputs.

{
  "documentId": "01H...",
  "text": "Anchor paragraph."
}
{
  "documentId": "01H...",
  "id": "p-AAAAAAAAAAAAAAAAAAAAAA",
  "text": "Updated text"
}
{
  "documentId": "01H...",
  "anchorId": "p-AAAAAAAAAAAAAAAAAAAAAA",
  "edge": "end",
  "rows": [
    ["A1", "B1"],
    ["A2", "B2"]
  ]
}

The first payload above is a complete find_elements_by_text call. The second is a replace_paragraph_text call. The third is an insert_table call.

Successful edit responses include summary plus machine-readable metadata: affectedElementIds, idMap, and invalidatedSelectors. If Nudocs returns a conflict (EDIT_CONFLICT), the MCP error payload includes structured details such as latest version/revision hints.

Supported File Formats

Input formats (upload):

  • Markdown (.md)
  • HTML (.html, .xhtml)
  • LaTeX (.latex, .tex)
  • reStructuredText (.rst)
  • Org mode (.org)
  • Textile (.textile)
  • DocBook XML
  • EPUB (.epub)
  • MediaWiki (.wiki)
  • Jupyter Notebook (.ipynb)
  • OpenDocument Text (.odt)
  • Microsoft Word (.doc, .docx)
  • Rich Text Format (.rtf)
  • Plain Text (.txt)
  • PDF (.pdf)

Output formats (download):

  • Markdown (.md)
  • HTML (.html, .xhtml)
  • LaTeX (.latex, .tex)
  • PDF (.pdf)
  • reStructuredText (.rst)
  • Org mode (.org)
  • Textile (.textile)
  • DocBook XML
  • EPUB (.epub)
  • Microsoft Word (.doc, .docx)
  • OpenDocument Text (.odt)
  • Rich Text Format (.rtf)
  • Plain Text (.txt)
  • MediaWiki (.wiki)
  • AsciiDoc (.adoc, .asciidoc)
  • Jupyter Notebook (.ipynb)

Prerequisites

Configuration

Environment Variables

Set your Nudocs API key:

export NUDOCS_API_KEY=your_api_key_here

You can find your API key in your Nudocs account settings.

Integration

Claude Desktop

Add this server to your Claude Desktop configuration file:

macOS: ~/Library/Application Support/Claude/claude_desktop_config.json Windows: %APPDATA%\Claude\claude_desktop_config.json

{
  "mcpServers": {
    "nudocs": {
      "command": "npx",
      "args": ["-y", "@nutrient-sdk/nudocs-mcp-server"],
      "env": {
        "NUDOCS_API_KEY": "your_api_key_here"
      }
    }
  }
}

After updating the configuration, completely quit and restart Claude Desktop. You'll see an MCP indicator in the bottom-right corner of the chat input when successfully connected.

Claude Code

Claude Code supports three configuration levels:

Local Configuration (Project-specific)

Create .mcp.json in your project root:

{
  "mcpServers": {
    "nudocs": {
      "command": "npx",
      "args": ["-y", "@nutrient-sdk/nudocs-mcp-server"],
      "env": {
        "NUDOCS_API_KEY": "your_api_key_here"
      }
    }
  }
}

Global Configuration (All projects)

Create or edit ~/.claude.json:

{
  "mcpServers": {
    "nudocs": {
      "command": "npx",
      "args": ["-y", "@nutrient-sdk/nudocs-mcp-server"],
      "env": {
        "NUDOCS_API_KEY": "your_api_key_here"
      }
    }
  }
}

Using the CLI

You can also use Claude Code's built-in configuration wizard:

claude mcp add --transport stdio nudocs --env NUDOCS_API_KEY=YOUR_KEY -- npx -y @nutrient-sdk/nudocs-mcp-server

OpenAI Codex

OpenAI Codex supports MCP servers through its CLI and IDE extension.

Using the CLI

The easiest way to add the Nudocs MCP server:

codex mcp add nudocs --env NUDOCS_API_KEY=YOUR_KEY -- npx -y @nutrient-sdk/nudocs-mcp-server

Manual Configuration

Edit ~/.codex/config.toml to add the server:

[mcp.nudocs]
command = "npx"
args = ["-y", "@nutrient-sdk/nudocs-mcp-server"]

[mcp.nudocs.env]
NUDOCS_API_KEY = "your_api_key_here"

The configuration file is shared between the Codex CLI and IDE extension, so you only need to configure it once.

For more information, visit the OpenAI Codex MCP documentation.

Development

Building from Source

npm install
npm run build

Running Locally

npm run dev

Unit Tests

npm run test:unit

Testing with MCP Inspector

The MCP Inspector allows you to test the server interactively:

npm run inspector

Issues & Support

Report issues at: https://github.com/PSPDFKit/nudocs-mcp-server/issues

License

See LICENSE file for details.


Made by Nutrient | Powered by Nudocs.ai

推荐服务器

Baidu Map

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。

官方
精选
JavaScript
Playwright MCP Server

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。

官方
精选
TypeScript
Audiense Insights MCP Server

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。

官方
精选
本地
TypeScript
Magic Component Platform (MCP)

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。

官方
精选
本地
TypeScript
VeyraX

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。

官方
精选
本地
Kagi MCP Server

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。

官方
精选
Python
graphlit-mcp-server

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。

官方
精选
TypeScript
Neon MCP Server

Neon MCP Server

用于与 Neon 管理 API 和数据库交互的 MCP 服务器

官方
精选
Exa MCP Server

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。

官方
精选
mcp-server-qdrant

mcp-server-qdrant

这个仓库展示了如何为向量搜索引擎 Qdrant 创建一个 MCP (Managed Control Plane) 服务器的示例。

官方
精选