
Parseflow
- 2 repo stars
- Updated December 18, 2025
- Libres-coder/ParseFlow
ParseFlow is an MCP server that extracts text, metadata, images, TOC, and search results from PDFs for your coding agent.
About
ParseFlow is an MCP server that exposes PDF parsing to coding agents over stdio. developers shipping RAG pipelines, contract review helpers, or doc-aware features can register parseflow-mcp-server so Claude Code or Cursor pulls clean text, page-level metadata, TOC structure, embedded images, and search hits instead of asking users to paste fragments. The server targets developers who already run Node-based MCP configs and want a narrow, dependable PDF layer rather than a full OCR platform. It fits when your agent must read uploaded PDFs during implementation—onboarding packs, API PDFs, or competitor one-pagers—without bolting on a separate SaaS viewer. Complexity is moderate: you install the npm package, add the server to your MCP manifest, and invoke tools from the agent thread. It does not replace hosted vector DBs or layout-heavy table reconstruction; it optimizes for extraction and discovery inside standard PDFs through the Model Context Protocol.
- Full-text extraction and in-document search over PDFs via MCP tools
- Metadata and table-of-contents access for structured navigation
- Image extraction from pages for multimodal or asset workflows
- stdio npm package parseflow-mcp-server (v1.0.2) for local agent attach
- MCP integration for agents, not a standalone PDF desktop app
Parseflow by the numbers
- Data as of Jul 7, 2026 (Skillselion catalog sync)
claude mcp add parseflow-mcp-server -- npx -y parseflow-mcp-serverAdd your badge
Show developers this MCP server is listed on Skillselion. Paste this into your README.
| repo stars | ★ 2 |
|---|---|
| Package | parseflow-mcp-server |
| Transport | STDIO |
| Auth | None |
| Last updated | December 18, 2025 |
| Repository | Libres-coder/ParseFlow ↗ |
What it does
Let Claude Code extract text, metadata, images, TOC, and search inside PDFs without manual copy-paste.
Who is it for?
Best when you're adding document-aware agents, RAG ingest, or internal tools that must consume PDF specs and reports.
Skip if: Skip if you need enterprise OCR, redaction workflows, or batch conversion outside an agent IDE session.
What you get
Your agent can query PDF contents, structure, and images in-thread so features and summaries stay grounded in the actual file.
- Extracted PDF text and metadata usable in agent prompts
- TOC and image payloads for downstream chunking or UI
- In-PDF search results without manual page hunting
By the numbers
- Server version 1.0.2 on npm identifier parseflow-mcp-server
- stdio transport via Model Context Protocol
- Capabilities: text extraction, metadata, search, images, and TOC
README.md
📄 ParseFlow
ParseFlow 是一个全面的文档解析解决方案,支持 PDF、Word、Excel、PowerPoint 和 图片 OCR。它提供独立的核心库和 MCP 服务器,可供 AI 助手使用。
✨ 功能特性
📄 PDF 支持
- ✅ 多策略文本提取(原始、格式化、清理)
- ✅ 按页或按范围提取
- ✅ 🔐 加密 PDF 密码支持
- ✅ 📄 PDF 合并、拆分、提取页面
- ✅ 元数据获取、全文搜索
📝 Word / 📊 Excel / 🎯 PowerPoint
- ✅ 文本提取和搜索
- ✅ HTML 转换(Word)
- ✅ 多工作表支持(Excel)
- ✅ 幻灯片提取(PowerPoint)
🔍 OCR 图片识别
- ✅ 支持 12 种语言
- ✅ 图片文字提取和搜索
🧠 语义搜索
- ✅ AI 向量嵌入
- ✅ 智能文档搜索(无需精确关键词)
📦 批量处理
- ✅ 并行处理多个文件
- ✅ 目录递归扫描
- ✅ 批量提取和搜索
🤖 MCP 服务器
- ✅ 20 个 AI 助手工具
- ✅ 支持 Claude Desktop、Windsurf、Cursor
📦 安装
核心库
npm install parseflow-core
MCP 服务器
npm install -g parseflow-mcp-server
# 或使用 npx
npx parseflow-mcp-server
🚀 快速开始
PDF 解析
import { PDFParser } from 'parseflow-core';
const parser = new PDFParser();
const text = await parser.extractText('document.pdf');
const results = await parser.search('document.pdf', '关键词');
Word 解析
import { WordParser } from 'parseflow-core';
const parser = new WordParser();
const result = await parser.extractText('report.docx');
const html = await parser.extractHTML('report.docx');
Excel 解析
import { ExcelParser } from 'parseflow-core';
const parser = new ExcelParser();
const data = await parser.extractData('spreadsheet.xlsx');
const results = await parser.searchText('data.xlsx', '收入');
PowerPoint 解析
import { PowerPointParser } from 'parseflow-core';
const parser = new PowerPointParser();
const result = await parser.extractText('presentation.pptx');
const results = await parser.searchText('slides.pptx', '关键词');
🛠️ MCP 服务器配置
Claude Desktop
在 claude_desktop_config.json 中添加:
{
"mcpServers": {
"parseflow": {
"command": "npx",
"args": ["-y", "parseflow-mcp-server"]
}
}
}
可用工具(23 个)
| 类别 | 工具 | 描述 |
|---|---|---|
extract_text |
提取文本(支持加密 PDF) | |
get_metadata |
获取元数据 | |
search_pdf |
全文搜索 | |
extract_images |
提取图片 | |
get_toc |
获取目录 | |
merge_pdf |
合并多个 PDF | |
split_pdf |
拆分为单页 | |
extract_pdf_pages |
提取指定页码 | |
add_watermark |
添加文字水印 | |
add_image_watermark |
添加图片水印 | |
remove_watermark |
移除水印(覆盖) | |
| Word | extract_word |
提取文本/HTML |
search_word |
文本搜索 | |
| Excel | extract_excel |
提取数据 |
search_excel |
单元格搜索 | |
| PPT | extract_powerpoint |
提取幻灯片 |
search_powerpoint |
幻灯片搜索 | |
| OCR | extract_ocr |
图片文字识别 |
search_ocr |
OCR 文本搜索 | |
| AI | semantic_index |
文档向量索引 |
semantic_search |
语义相似搜索 | |
| 批量 | batch_extract |
批量提取多文件 |
batch_search |
批量搜索多文件 |
📈 版本历史
| 版本 | 功能 |
|---|---|
| v1.8.0 | 💧 PDF 水印(文字/图片水印) |
| v1.7.0 | 📦 批量处理(并行处理多文件) |
| v1.6.0 | 🧠 语义搜索(AI 向量嵌入) |
| v1.5.0 | 📄 PDF 合并/拆分/提取 |
| v1.4.0 | 🔐 加密 PDF 支持 |
| v1.3.0 | 🔍 OCR 图片文字识别 |
| v1.2.0 | 🎯 PowerPoint 支持 |
| v1.1.0 | 📝 Word + 📊 Excel 支持 |
| v1.0.0 | 📄 PDF 基础解析 |
🔗 链接
- npm Core: https://www.npmjs.com/package/parseflow-core
- npm MCP: https://www.npmjs.com/package/parseflow-mcp-server
- GitHub: https://github.com/Libres-coder/ParseFlow
📄 许可证
MIT License - 详见 LICENSE
Made with ❤️ by Libres-coder
Recommended MCP Servers
How it compares
MCP PDF toolkit, not a hosted document SaaS or a general web-scraping skill.
FAQ
Who is ParseFlow for?
Developers and agent developers who want PDF text, metadata, TOC, images, and search exposed as MCP tools inside Claude Code, Cursor, or similar hosts.
When should I use ParseFlow?
Use it during build when your product or agent workflow must read, search, or chunk PDFs programmatically instead of manual export.
How do I add ParseFlow to my agent?
Install the npm package parseflow-mcp-server, register it as a stdio MCP server in your client config, then call its tools from the agent session.