Intelligent Document Extraction - ComIDP

MCP设计创作

提供方 ComPDFKit

设计创作 未标注 API Key 社区 / 第三方 Claude DesktopClaudeCursorGemini CLI
查看接入方式

ComIDP MCP Server is a lightweight Model Context Protocol (MCP) server designed for seamless integrating ComIDP with AI chatbots, providing unstructured document processing functionalities, such as extracting data from PDF files. The service returns results in structured plain-te

一句话选型结论

按资料字段整理

Intelligent Document Extraction - ComIDP 属于设计创作类 MCP 服务器,传输方式未标注,API Key,可直接接入 Claude Desktop、Claude、Cursor 等,有可直接执行的安装命令。接入前先确认权限范围;涉及写入、删除或付款的操作,建议先在测试环境验证。

分类设计创作 传输方式未标注 鉴权API Key 部署方式未标注
GitHub 星标 未标注 仓库热度,无公开仓库时显示未标注
工具数 未标注 暴露给模型的工具数量
资源 / 提示词 0 / 0 资源与提示词模板数量
维护方 社区 / 第三方 核验等级 仅资料核对

工具与能力

设计创作
数量来自来源目录的标注,具体工具清单以官方文档为准
工具数量未标注暴露给模型可调用的工具个数
资源数量0可读取的资源条目
提示词数量0内置提示词模板数量

能力标签

支持客户端

适用场景

ai agentdata extractionai assistant

安装与配置

未标注 API Key
上游未给出传输方式
pip install uv \r

这条命令只是入口,具体启动参数请按官方文档补齐,不要直接照搬到别的服务器。

  1. 1
    准备运行环境

    按官方文档确认运行环境(常见为 Node.js 或 Python),装好后再执行安装命令。

  2. 2
    安装服务

    复制上面的安装命令执行;命令里带占位符的位置,换成自己的密钥或路径。

  3. 3
    写入客户端配置

    在客户端的 MCP 配置里填好启动命令和参数,保存后完全退出客户端再打开,让新配置生效。

  4. 4
    验证是否可用

    连接成功后随便调一个只读工具,确认能返回结果,再接入正式流程。

典型使用场景

3 类
按分类与标签整理的常见用法,具体以官方文档为准
出图与素材

生成配图、图标与素材。

设计稿处理

读取设计稿、批量改尺寸或导出。

内容生成

按主题生成文案与视觉草稿。

选择建议

选之前先看这几条
按条目字段给出的判断,不替你决定用不用

值得选的地方

  • 有现成的安装命令,不用自己找入口,照着跑就能起服务。
  • 已适配 Claude Desktop、Claude、Cursor,换客户端不用重新研究配置。

先想清楚的地方

  • 使用前要准备好API Key对应的凭据,并按最小权限配置。
  • 资料没标注工具数量,实际能力要装好后确认。
  • 版本更新与行为细节以官方仓库为准。

更适合

  • 需要在设计创作场景里让模型直接调用外部能力的人。
  • 已经在用 Claude Desktop、Claude 的用户,接入路径最短。
  • 想先快速试一下、能接受在本地或自有环境跑服务的人。

不太适合

  • 不方便管理密钥或第三方授权的人。
  • 需要厂商级 SLA 和技术支持的生产环境。

同类对比

6 个同类
同分类里挑最接近的几条,字段以各自详情页为准
MCP 服务器工具数传输方式部署方式安装命令最后核验
Intelligent Document Extraction - ComIDP当前 未标注 未标注 未标注 有 未标注
Cursor Talk To Figma MCP Da1c7911 40 未标注 未标注 无 2026-05-17
Blender MCP 17 stdio · 本地进程 本地部署 有 2026-05-18
Nova Canvas MCP Server 2 stdio · 本地进程 本地部署 有 2026-05-17
Figma Context MCP 3aaa8d28 未标注 未标注 支持远程部署 无 2026-05-17
kreuzberg 未标注 stdio · 本地进程 本地部署 有 2026-04-12

服务器档案

设计创作
字段来自项目仓库与公开索引,中文说明由本站标注
提供方ComPDFKit上游标注的维护方 ComPDFKit
分类设计创作本站按上游分类展示
连接方式未标注上游未给出传输方式
鉴权方式API Key远程服务建议用最小权限凭据
维护方社区 / 第三方核验等级 仅资料核对
包名 / 运行时未标注 · Cross-platform本地安装时用到的包标识与运行时
部署类型未标注上游给出的部署形态
最近核验未标注上游最后核验时间

详细介绍

19.3 KB
来自项目仓库与上游目录的原文,未做删改

ComPDF MCP

ComPDF MCP is built for MCP clients and AI agent platforms. It provides PDF/image parsing and data extraction, PDF conversion, PDF editing, and image conversion capabilities. It helps users process raw documents first, then pass lighter, cleaner, and more structured results into AI, reducing token usage, lowering AI costs, and improving overall processing efficiency.

> * If you find ComPDF MCP useful, please consider giving us a ⭐ Star on GitHub. It helps us grow and improve.
> * Got questions or ideas? Join the conversation in our Discussions.

Why Choose ComPDF MCP •
Capabilities •
Technical Architecture •
Quick Start •
Deployment and Operations •
License and Free Access •
Use Cases and Example Prompts •
Support

Why Choose ComPDF MCP

The core reason users choose ComPDF MCP is not simply to let AI work with PDFs. It is to process complex, lengthy, and noisy source documents into inputs that are easier for models to understand before handing them off to AI for analysis, summarization, Q&A, or automation.

Raw PDFs, scanned files, and image files are often larger, noisier, and less stable in structure. By using ComPDF MCP for conversion, extraction, and page-level processing first, then sending the results into AI, users can:

  • Reduce AI token usage
  • Lower AI costs
  • Shorten response time
  • Improve the quality of structured outputs
  • Avoid wasting model budget on noisy raw files
  • Handle both PDF and image processing through one MCP service
  • Fit document-heavy workflows and multi-platform agent scenarios

This approach is especially valuable when processing reports, contracts, invoices, tables, scanned files, and similar documents in bulk. In these cases, a "process documents first, then call AI" workflow is usually more cost-efficient and better suited for downstream automation than sending original files directly to the model.

Capabilities

ComPDF MCP focuses on three major capability groups: document conversion, PDF operations, and intelligent parsing and data extraction.

1. PDF and Image Conversion

Capability Description
PDF to Word Convert PDF files into editable Word documents while preserving the original layout, text, images, and formatting as much as possible.
PDF to Excel Convert PDF files into Excel workbooks with support for tables, numbers, and structured business data.
PDF to Slide Convert PDF pages into editable PowerPoint slides while preserving the original layout and visual structure as much as possible.
PDF to HTML Convert PDF files into HTML for web display and content reuse while retaining text, images, tables, and layout.
PDF to RTF Convert PDF files into RTF documents with support for text and image content.
PDF to Image Convert PDF pages into PNG or JPG images with configurable resolution and DPI.
PDF to CSV Extract tables from PDF files and export them as CSV, either table by table or as merged output.
PDF to TXT Extract text from PDF or scanned PDF files and save it as plain text.
PDF to JSON Extract text, tables, and images from PDF files and save them as structured JSON.
PDF to Markdown Convert PDF files into Markdown for easier reuse in knowledge bases, developer docs, blog systems, and AI workflows.
PDF to Searchable PDF Run OCR on scanned PDFs and output searchable PDFs with selectable, copyable, and highlightable text for retrieval, archiving, and downstream processing.
PDF to OFD Convert PDF files into OFD files for OFD archiving, document circulation, and localized office scenarios.
Word to PDF Convert Word documents into PDF while preserving layout, fonts, images, and page structure as much as possible for sharing, archiving, and printing.
PNG to PDF Convert PNG images into PDF for easier packaging, sharing, printing, and archiving of screenshots, design files, or supporting image materials.
RTF to PDF Convert RTF documents into PDF while preserving core text styling and layout for consistent cross-device viewing and formal output.
Excel to PDF Convert Excel workbooks or spreadsheets into PDF for report sharing, printing, archiving, and preventing accidental formula edits.
TXT to PDF Convert plain TXT files into PDF, making logs, notes, and text instructions easier to organize as fixed-layout documents.
CSV to PDF Convert CSV table data into PDF for snapshot sharing, review, printing, and business archiving.
Slide to PDF Convert PowerPoint presentations into PDF for presentation distribution, cross-device viewing, and formal record keeping.
HTML to PDF Convert HTML pages or content fragments into PDF for webpage preservation, report export, email archiving, and printable output.
Image to Word Convert JPG, JPEG, PNG, and BMP image files into editable Word documents.
Image to Excel Convert image files into Excel workbooks with support for tables, text, and numeric content.
Image to PPT Convert image files into editable PowerPoint slides while preserving visible layout and content structure as much as possible.
Image to PDF Convert JPG, JPEG, PNG, BMP, and similar image files into PDF for consolidating, sharing, printing, and archiving one or multiple images.
Image to HTML Convert image files into HTML while preserving text, layout, tables, and major visual elements as much as possible.
Image to RTF Convert image files into RTF documents with support for extracted text and images.
Image to CSV Extract tables from image files and export them as CSV.
Image to TXT Extract text from image files and save it as plain text.
Image to JSON Extract text, tables, and images from image files and save them as structured JSON.

2. PDF Editing, Protection, and Comparison

Capability Description
Merge PDF Files Combine multiple PDF files into a single PDF document.
Split PDF Files Split one PDF file into multiple smaller PDF files.
Rotate PDF Pages Rotate selected PDF pages by 90, 180, or 270 degrees.
Delete PDF Pages Remove one or more pages from a PDF file.
Insert PDF Pages Insert pages from another PDF into an existing PDF.
Extract PDF Pages Extract selected pages or page ranges and save them as a new file.
Convert PDF Standard Convert a PDF conformance or archival standard.
Add Watermark Add text or image watermarks to PDF files for branding or usage control.
Remove Watermark Remove text or image watermarks from supported PDF files.
Compress PDF Reduce PDF file size for easier storage, upload, and sharing.
Encrypt PDF Protect PDF files with AES encryption and permission controls.
Decrypt PDF Remove passwords from authorized PDF files for internal processing or reuse.
Compare PDFs Compare the content differences between two PDF files.

3. Intelligent Parsing and Data Extraction

Capability Description
Intelligent Document Parsing Parse PDFs and images into structured document output for agents, automation workflows, and downstream systems.
Intelligent Data Extraction from PDFs and Images Extract valuable business data such as text, tables, and content fields from PDFs and images.

Technical Architecture

This project is an MCP Streamable HTTP service built with Python, FastMCP, and Starlette. Tool parameter schemas are generated from the bundled official OpenAPI parameter snapshot, and the service rejects unknown parameters.

Clients can access every tool through the aggregate endpoint:```text
http://127.0.0.1:8000/mcp


If you only need part of ComPDF MCP, use the module endpoints below:

| Module                                  | MCP Endpoint          | Business Tools | Representative Tools                                                       |
| --------------------------------------- | --------------------- | --------------:| -------------------------------------------------------------------------- |
| PDF and Image Conversion                | `/mcp/conversion/mcp` | 28             | `pdf_to_word`, `pdf_to_markdown`, `image_to_json`, `word_to_pdf`, and more |
| Intelligent Parsing and Data Extraction | `/mcp/ai/mcp`         | 2              | `document_parse`, `document_extract`                                       |
| PDF Editing, Protection, and Comparison | `/mcp/pdf/mcp`        | 13             | `merge_pdf`, `add_watermark`, `encrypt_pdf`, `compare_pdf`, and more       |
| All capabilities                        | `/mcp`                | 44             | All tools above                                                            |

All business tools use a shared file-input shape. Keys in `options` must use the camelCase parameter names from the ComPDF API documentation. Only `htmlFile`, `templateFile`, `dataFile`, `imageFile`, and `iccFile` may be uploaded through `special_files`.```json
{
  "files": [
    {
      "filename": "report.pdf",
      "content_base64": "JVBERi0xLjc...",
      "content_type": "application/pdf"
    }
  ],
  "options": {
    "pageRanges": "1-3,6",
    "enableOcr": 1
  },
  "special_files": {
    "imageFile": {
      "filename": "watermark.png",
      "content_base64": "iVBORw0..."
    }
  }
}

The decoded size of an individual Base64 file is limited to 100 MB. Synchronous tool responses include the ComPDF result, the relevant official documentation URL, and a statement that download URLs expire after 24 hours.

Quick Start

1. Install and Run the Service

Prerequisite: Python 3.10+. From the repository root, create a virtual environment and install the package in editable mode.

Windows:```powershell
python -m venv .venv
.venv\Scripts\Activate.ps1
pip install -e .
Copy-Item .env.example .env
compdf-streaming-mcp


Linux / macOS:```bash
python3 -m venv .venv
source .venv/bin/activate
pip install -e .
cp .env.example .env
compdf-streaming-mcp

The service listens on 0.0.0.0:8000 by default. Use these endpoints to check its status:

  • GET /healthz: service status and mounted MCP routes.
  • GET /readyz: readiness status without calling the upstream ComPDF API.
  • GET /metrics: request counters in Prometheus text format.

2. Configure the Service

Copy .env.example to .env, then adjust settings for the deployment environment. The server does not store or configure a user's ComPDF API key. Each MCP request supplies that user's key through the X-ComPDF-API-Key header. You can sign up for ComPDF Portal and copy your free API key from the location shown below.```dotenv
COMPDF_API_BASE_URL=https://api-server.compdf.com/server
COMPDF_API_TIMEOUT_SECONDS=180
MCP_HOST=0.0.0.0
MCP_PORT=8000
MCP_ALLOWED_HOSTS=localhost,localhost:,127.0.0.1,127.0.0.1:,mcp.example.com
MCP_RATE_LIMIT_PER_MINUTE=120


For mainland China accounts, use `COMPDF_API_BASE_URL=https://api-server.compdf.cn/server`. In production, add the reverse-proxy domain to `MCP_ALLOWED_HOSTS`; include a non-default port or use `mcp.example.com:*` when appropriate.

### 3. Configure an MCP Client

The examples below use the unified MCP URL `http://127.0.0.1:8000/mcp`. For a remote deployment, replace it with the public HTTPS address, for example `https://your-domain.example/mcp`.

First obtain a personal ComPDF API key (sign up to get your free API key). For clients that use environment variables, set it on the client machine first:```powershell
$env:COMPDF_API_KEY = "your-own-compdf-api-key"

Codex

Add the remote MCP service to ~/.codex/config.toml:```toml
[mcp_servers.compdf]
url = "http://127.0.0.1:8000/mcp"
env_http_headers = { "X-ComPDF-API-Key" = "COMPDF_API_KEY" }
tool_timeout_sec = 180


#### Claude (Claude Code)

Add the following MCP server to the user-level configuration file `~/.claude.json`. `COMPDF_API_KEY` must be available in the terminal session used to start Claude Code:```json
{
  "mcpServers": {
    "compdf": {
      "type": "http",
      "url": "http://127.0.0.1:8000/mcp",
      "headers": {
        "X-ComPDF-API-Key": "${COMPDF_API_KEY}"
      }
    }
  }
}

If ~/.claude.json already has configuration, merge compdf into the existing mcpServers object instead of overwriting the file.

GitHub Copilot (VS Code)

Run MCP: Open User Configuration from the VS Code Command Palette, or create .vscode/mcp.json in the workspace, then add the following configuration. VS Code prompts for the API key in a password field on first use:```json
{
"inputs": [
{
"type": "promptString",
"id": "compdf-api-key",
"description": "ComPDF API Key",
"password": true
}
],
"servers": {
"compdf": {
"type": "http",
"url": "http://127.0.0.1:8000/mcp",
"headers": {
"X-ComPDF-API-Key": "${input:compdf-api-key}"
}
}
}
}


#### Cursor

Open **Cursor Settings > Tools & MCP**, select **New MCP Server**, or edit `~/.cursor/mcp.json`, then add:```json
{
  "mcpServers": {
    "compdf": {
      "url": "http://127.0.0.1:8000/mcp",
      "headers": {
        "X-ComPDF-API-Key": "your-own-compdf-api-key"
      }
    }
  }
}

This Cursor configuration stores the key in a local configuration file. Do not commit ~/.cursor/mcp.json or a project-level configuration file that contains the key.

All clients must send X-ComPDF-API-Key with every MCP request. Do not pass the key as a tool argument, URL parameter, or chat message.

4. Asynchronous and Presigned Uploads

In addition to the 44 synchronous business tools, each module provides list_operations and these task tools:

  • start_async_operation: invokes the matching /v2/processAsync/... endpoint for long-running or multi-file work.
  • get_task_status: queries task status from /v2/task/taskInfo.
  • create_presigned_upload, upload_presigned_file, and start_presigned_operation: create a presigned task, upload the file, and start the task in sequence.

Presigned uploads support only operations with one standard file. Use asynchronous mode for PDF merging, comparison, page inserting, and more. The presigned URL remains only in the creating module server's memory and is never exposed as a tool argument or response field.

Deployment and Operations

Build and run the service with Docker Compose:powershell docker compose up --build

In production, expose the aggregate endpoint as https://your-domain.example/mcp behind an HTTPS reverse proxy, and configure MCP_ENV=production, MCP_PUBLIC_URL, MCP_OAUTH_ISSUER_URL, and MCP_STATIC_TOKENS_JSON. When enabled, clients must also send Authorization: Bearer; this service token is separate from each user's ComPDF API key. See DEPLOYMENT.md for image transfer, HTTPS proxy, update, and rollback instructions.

License and Free Access

ComPDF MCP can be used for free, without purchasing or contacting sales.

  • Sign up to get your free API key and configure it

  • Suitable for individual trials, feature validation, and workflow testing

  • If you need higher API assets, enterprise deployment, or commercial cooperation, please contact the sales

This lowers the barrier to adoption and makes it easier to test real AI document workflows before deciding whether to move into deeper business use.

Use Cases and Example Prompts

Upload a PDF, image, or other source file. Enter a task instruction such as extracting tables, converting formats, merging PDFs, or adding watermarks. The agent calls the corresponding ComPDF Server tool and returns the result. If deeper analysis is needed, pass the processed output to AI afterward.

Example scenarios:

  • In Claude, Cursor, Cline, and similar MCP clients, users upload reports, manuals, or proposal PDFs, convert them to Markdown or Word first, and then ask AI for summaries, knowledge Q&A, or content restructuring
  • In MCP workflows for invoices, statements, scanned tables, and image attachments, users extract tables and structured data first, then move into finance review, data entry, or automated routing
  • When organizing contracts, bids, quotations, or archived files in agent workflows, users first merge, split, watermark, and convert PDFs, then hand them off to AI for organization, naming, or outbound preparation
  • In multi-step automation workflows, users first turn PDF or image files into lightweight CSV, JSON, or TXT outputs, then let downstream agents handle field normalization, approval flows, knowledge base ingestion, or script orchestration

Example prompts:

  • Convert this PDF into Word and preserve the layout as much as possible.
  • Extract all tables from this PDF and export them as CSV.
  • Convert this image into JSON and return structured content.
  • Merge these PDFs, add a watermark, and return the final file.
  • Convert this report into Markdown first, then summarize the key points.

Built by the ComPDF team.

Website ·
Contact Sales ·
Tech Support

常见问题

7 问
答案由本站依据条目字段整理,不引用上游原文
Intelligent Document Extraction - ComIDP 怎么安装?

在客户端按下面的安装命令接入:pip install uv \r。装完在客户端配置里引用同一个包或端点。

Intelligent Document Extraction - ComIDP 怎么在 Claude Desktop、Claude、Cursor 里配置?

它的传输方式未标注,在客户端的 MCP 配置里填好启动命令和参数,保存后完全退出客户端再打开,配置才会生效。

Intelligent Document Extraction - ComIDP 收费吗?

本站不代收费用,也没有标注具体价格。是否收费取决于服务本身和背后的接口额度,正式使用前按官方说明确认。

Intelligent Document Extraction - ComIDP 需要 API Key 吗?

资料标注需要API Key认证,使用前准备好对应凭据,并按最小权限配置。

Intelligent Document Extraction - ComIDP 支持哪些客户端?

资料标注支持 Claude Desktop、Claude、Cursor、Gemini CLI、Codex、Zed;其他客户端只要能配置 MCP Server,也能用同样的命令或地址接入。

Intelligent Document Extraction - ComIDP 能提供哪些能力?

资料显示能力标签包括 ai agent、data extraction、ai assistant。具体工具清单以官方文档为准。

Intelligent Document Extraction - ComIDP 需要什么运行环境?

运行时为 Cross-platform。

热门专题

按标签挑服务器
全部专题 →

服务器信息以作者仓库与官方文档为准;分类、传输方式与鉴权方式的中文说明为中文整理,最近核验 未标注。