Sibyl 50762f7b

MCP文档知识

提供方 KyleBrandon

文档知识 未标注 无需鉴权 社区 / 第三方 Claude DesktopClaudeVS Code

Sibyl提供基于MCP协议的PDF处理和笔记管理服务,支持OCR、智能合并和模板创建等功能,适用于文档处理和知识管理场景。

一句话选型结论

按资料字段整理

Sibyl 50762f7b 属于文档知识类 MCP 服务器,传输方式未标注,无需鉴权,可直接接入 Claude Desktop、Claude、VS Code,没有现成安装命令,需按官方文档接入。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。

分类文档知识 传输方式未标注 鉴权无需鉴权 部署方式未标注
GitHub 星标 2 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

文档知识
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

PDF处理Go本地部署笔记管理OCR知识管理文档转换

安装与配置

未标注 无需鉴权
上游未给出传输方式
  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
文档检索问答

在你的文档库里按问题找答案。

知识库读写

把新资料写进知识库,随用随取。

内容整理

把零散资料归纳成结构化笔记。

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

  • 已适配 Claude Desktop、Claude、VS Code,换客户端不用重新研究配置。
  • 资料标注无需鉴权,装好就能连。

先想清楚的地方

  • 没有现成安装命令,需要按官方文档自己填启动命令和参数。
  • 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

  • 需要在文档知识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

  • 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
Sibyl 50762f7b当前 未标注 未标注 未标注 无 2026-05-17
dify 未标注 stdio · 本地进程 本地部署 有 2026-04-11
open-webui 未标注 stdio · 本地进程 本地部署 有 2026-04-11
JeecgBoot 未标注 stdio · 本地进程 本地部署 有 2026-04-11
Cherry Studio 未标注 未标注 未标注 无 2026-05-17
mempalace 未标注 stdio · 本地进程 本地部署 有 2026-04-11

服务器档案

文档知识
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KyleBrandon上游标注的维护方 KyleBrandon
分类文档知识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
鉴权方式无需鉴权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 · 未标注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2026-05-17上游最后核验时间

详细介绍

17.0 KB
来自项目仓库与上游目录的原文,未做删改

Sibyl MCP Servers

AI-powered PDF processing and note management through the Model Context Protocol (MCP)

Sibyl provides two specialized MCP servers that enable LLM hosts (Claude Desktop, VS Code, etc.) to intelligently process PDFs and manage markdown notes with advanced capabilities like OCR, intelligent merging, and template-based creation.

⚖️ Important: Licensing Notice

Before you begin: Sibyl uses MuPDF (via go-fitz) for PDF processing, which is licensed under AGPL v3. This means:

  • ✅ Open source projects: Free to use under AGPL v3
  • ⚠️ Commercial/proprietary use: Requires commercial license from Artifex Software
  • 📋 Network deployment: Must provide source code to users under AGPL v3

See detailed licensing information below before deploying in production.

🚀 Quick Start

Prerequisites

  • Go 1.21+ - Install Go
  • Google Cloud Account - For Google Drive API access
  • Mathpix Account - For OCR processing (Sign up)
  • MCP-compatible host - Claude Desktop, VS Code with MCP extension, etc.

1. Clone and Build```bash

git clone https://github.com/your-username/sibyl.git
cd sibyl
make all


This creates:
- `./bin/pdf-server` - PDF processing MCP server
- `./bin/notes-server` - Note management MCP server

### 2. Set Up Credentials

#### Google Drive API Setup
1. Go to Google Cloud Console
2. Create a new project or select existing one
3. Enable the Google Drive API
4. Create service account credentials
5. Download the JSON credentials file
6. Share your Google Drive folder with the service account email

#### Mathpix OCR Setup
1. Sign up at Mathpix
2. Create a new app in your dashboard
3. Note your App ID and App Key

### 3. Configure Environment

Create a `.env` file in the project root:```bash
# Google Drive Configuration
GOOGLE_APPLICATION_CREDENTIALS=/path/to/your/google-credentials.json
GCP_FOLDER_ID=your-google-drive-folder-id

# Mathpix OCR Configuration
MATHPIX_APP_ID=your-mathpix-app-id
MATHPIX_APP_KEY=your-mathpix-app-key

4. Configure Your MCP Host

Add the servers to your MCP host configuration:

Claude Desktop (~/Library/Application Support/Claude/claude_desktop_config.json on macOS):```json
{
"mcpServers": {
"pdf-server": {
"command": "/full/path/to/sibyl/bin/pdf-server",
"args": [
"--credentials", "/path/to/google-credentials.json",
"--folder-id", "your-google-drive-folder-id",
"--mathpix-app-id", "your-mathpix-app-id",
"--mathpix-app-key", "your-mathpix-app-key",
"--ocr-languages", "en,fr,de",
"--log-level", "INFO"
]
},
"notes-server": {
"command": "/full/path/to/sibyl/bin/notes-server",
"args": [
"--notesFolder", "/path/to/your/notes",
"--logLevel", "INFO",
"--logFile", "notes-server.log"
]
}
}
}


### 5. Test the Setup

1. Restart your MCP host (Claude Desktop, VS Code, etc.)
2. Try asking: *"What PDF documents do I have available?"*
3. Try asking: *"Show me my notes in the projects folder"*

## 📖 Core Concepts

### PDF Server Capabilities

The PDF server connects to your Google Drive and provides:

- **🔍 Document Search**: Find PDFs by name, content, or metadata
- **📄 PDF-to-Markdown Conversion**: High-quality conversion using Mathpix OCR + visual analysis
- **🖼️ Image Processing**: Converts PDFs to 150 DPI PNG images for visual analysis
- **📊 MCP Resources**: Structured access to your document library

**How it works:**
1. **PDF → PNG**: Converts PDF pages to high-quality images using MuPDF
2. **PDF → OCR**: Extracts text using Mathpix OCR (handles formulas, tables, handwriting)
3. **PNG + OCR → LLM**: Your LLM analyzes both visual and text data for optimal conversion

### Notes Server Capabilities

The notes server manages your markdown vault with:

- **📝 Intelligent Reading/Writing**: Full CRUD operations on markdown files
- **🔄 Smart Merging**: 5 merge strategies (append, prepend, date_section, topic_merge, replace)
- **👀 Merge Preview**: See exactly what changes before applying them
- **📋 Template System**: Pre-built templates for daily notes, meetings, research, projects
- **🔍 Content Search**: Full-text search across your entire vault
- **📊 MCP Resources**: Structured exploration of your note collection

## 🛠️ Available Tools

### PDF Server Tools

| Tool | Description | Parameters |
|------|-------------|------------|
| `search_pdfs` | Search Google Drive for PDF files | `query` (string), `max_files` (number) |
| `convert_pdf_to_markdown` | Convert PDF to Markdown using OCR + visual analysis | `file_id` (string) |

### Notes Server Tools

| Tool | Description | Parameters |
|------|-------------|------------|
| `read_note` | Read note content | `path` (string) |
| `write_note` | Create or overwrite a note | `path` (string), `content` (string) |
| `merge_note` | Merge content with existing note | `path`, `content`, `strategy`, `title?` |
| `preview_merge` | Preview merge operation | `path`, `content`, `strategy?` |
| `list_notes` | List notes in directory | `path?`, `recursive?` (boolean) |
| `search_notes` | Search note content | `query` (string), `path?` |
| `get_note_templates` | Get available templates | `template_type?` (string) |
| `create_note_from_template` | Create note from template | `path`, `template_type`, `variables?` |

### Merge Strategies

- **`append`** - Add content to end of file
- **`prepend`** - Add content to beginning of file
- **`date_section`** - Add as new dated section with timestamp
- **`topic_merge`** - Intelligent merging based on content topics
- **`replace`** - Replace entire file content

### Note Templates

- **`daily`** - Daily note with focus areas, tasks, and reflections
- **`meeting`** - Meeting notes with agenda, attendees, and action items
- **`research`** - Research notes with citations, analysis, and conclusions
- **`project`** - Project planning template with goals, milestones, and progress

## 🎯 Usage Examples

### Example 1: PDF Research Workflow

**You**: *"I have a research paper about machine learning in my Drive. Can you convert it to notes and add it to my research collection?"*

**Assistant**: I'll help you find and convert that research paper. Let me search your Google Drive first.

*→ Calls `search_pdfs(query: "machine learning")`*

I found 3 PDFs about machine learning:
1. "Deep Learning Fundamentals.pdf" (2.1MB, modified yesterday)
2. "ML Algorithms Survey.pdf" (896KB, modified last week)
3. "Neural Networks Intro.pdf" (1.5MB, modified last month)

Which would you like me to convert?

**You**: *"The first one"*

**Assistant**: I'll convert "Deep Learning Fundamentals.pdf" to Markdown using our OCR and visual analysis process.

*→ Calls `convert_pdf_to_markdown(file_id: "1BxYz...")`*

Perfect! The PDF has been processed with Mathpix OCR and converted to high-quality images. I've extracted the text and can see the visual elements including equations, diagrams, and tables.

Here's the converted content... [shows converted markdown]

Would you like me to save this as a new research note?

**You**: *"Yes, save it as 'Deep Learning Fundamentals Summary'"*

**Assistant**: *→ Calls `write_note(path: "research/Deep Learning Fundamentals Summary.md", content: converted_content)`*

Done! I've saved the converted paper as "Deep Learning Fundamentals Summary.md" in your research folder. The conversion includes all the mathematical formulas, diagrams, and structured content from the original PDF.

### Example 2: Smart Note Merging

**You**: *"I just took some meeting notes. Can you merge them with my daily note but keep things organized?"*

**Assistant**: I can help merge your meeting notes. Let me preview how this would look with topic-based merging to keep everything organized.

*→ Calls `preview_merge(path: "daily.md", content: meeting_notes, strategy: "topic_merge")`*

Here's how the merge would look:
- Your existing "Tasks" section would be combined with new action items
- Meeting notes would be added under a new "Meetings" section
- No duplicate content would be created

Should I proceed with this merge?

**You**: *"Perfect, go ahead"*

**Assistant**: *→ Calls `merge_note(path: "daily.md", content: meeting_notes, strategy: "topic_merge")`*

Done! Your meeting notes have been intelligently merged into your daily note with proper organization and no duplicates.

### Example 3: Template-Based Note Creation

**You**: *"Create a meeting note for tomorrow's sprint planning session"*

**Assistant**: I'll create a meeting note using our meeting template.

*→ Calls `create_note_from_template(path: "meetings/sprint-planning-2025-08-10.md", template_type: "meeting", variables: {"TITLE": "Sprint Planning", "DATE": "2025-08-10", "ATTENDEES": "TBD"})`*

I've created "sprint-planning-2025-08-10.md" with a structured meeting template including:
- Agenda section
- Attendees list (you can fill this in)
- Discussion points
- Action items tracking
- Next steps

The note is ready for you to add specific agenda items before the meeting.

## 📊 MCP Resources

Both servers provide MCP resources that enable intelligent exploration of your content without knowing specific file paths.

### PDF Server Resources

- **`pdf://documents/`** - Complete catalog of your PDF library with metadata
- **Example**: Lists all PDFs with file sizes, modification dates, and direct links

### Notes Server Resources

- **`notes://files/`** - Your complete note collection with previews and tags
- **`notes://templates/`** - Available note templates with descriptions
- **`notes://collections/`** - Notes organized by folders and tags

**Resource Benefits:**
- 🔍 **Discovery**: LLMs can explore without knowing file paths
- ⚡ **Efficiency**: Batch metadata retrieval vs individual queries
- 🧠 **Context**: Rich metadata helps LLMs make smarter decisions
- 🗺️ **Navigation**: Structured browsing of content hierarchies

## ⚙️ Configuration Options

### PDF Server Arguments```bash
./bin/pdf-server --help
Argument Required Description Environment Variable
--credentials Yes Path to Google Cloud credentials JSON GOOGLE_APPLICATION_CREDENTIALS
--folder-id Yes Google Drive folder ID to search GCP_FOLDER_ID
--mathpix-app-id Yes Mathpix API App ID MATHPIX_APP_ID
--mathpix-app-key Yes Mathpix API App Key MATHPIX_APP_KEY
--ocr-languages No Comma-separated language codes (default: "en") -
--log-level No Log level: DEBUG, INFO, WARN, ERROR (default: INFO) -
--log-file No Log file path (default: stderr) -

Notes Server Arguments```bash

./bin/notes-server --help


| Argument | Required | Description |
|----------|----------|-------------|
| `--notesFolder` | Yes | Path to your notes directory |
| `--logLevel` | No | Log level: DEBUG, INFO, WARN, ERROR (default: INFO) |
| `--logFile` | No | Log file path (default: stderr) |

## 🧪 Development & Testing

### Running Tests```bash
# Run all tests
go test ./...

# Run tests with coverage
go test -race -coverprofile=coverage.out -covermode=atomic ./pkg/...
go tool cover -html=coverage.out -o coverage.html

# Run specific package tests
go test ./pkg/pdfmcp
go test ./pkg/notes

# Run integration tests
go test ./tests/integration/...

Code Quality```bash

Static analysis

go vet ./...
staticcheck ./...

Security scanning

gosec ./...

Comprehensive linting

golangci-lint run --timeout=5m


### Project Structure```
sibyl/
├── cmd/                    # Main applications
│   ├── pdfserver/          # PDF MCP server entry point
│   └── noteserver/         # Notes MCP server entry point
├── pkg/                    # Reusable packages
│   ├── pdfmcp/             # PDF server implementation
│   ├── notes/              # Notes server implementation
│   ├── dto/                # Data transfer objects
│   └── utils/              # Shared utilities
├── tests/                  # Testing infrastructure
│   ├── integration/        # End-to-end tests
│   └── testutils/          # Test helpers
├── examples/               # Usage examples and configs
└── bin/                    # Built binaries

🔧 Troubleshooting

Common Issues

PDF Server won't start:
- ✅ Check Google Cloud credentials file exists and is valid
- ✅ Verify Google Drive API is enabled in your GCP project
- ✅ Ensure service account has access to your Drive folder
- ✅ Confirm Mathpix credentials are correct

Notes Server can't find notes:
- ✅ Verify notes folder path exists and is readable
- ✅ Check that notes are in Markdown format (.md extension)
- ✅ Ensure proper file permissions

MCP Host can't connect:
- ✅ Use absolute paths in MCP configuration
- ✅ Restart your MCP host after configuration changes
- ✅ Check server logs for startup errors

Log Files

Enable debug logging for troubleshooting:```bash

PDF Server

./bin/pdf-server --log-level DEBUG --log-file pdf-server.log [other args...]

Notes Server

./bin/notes-server --logLevel DEBUG --logFile notes-server.log [other args...]


## 📚 Advanced Usage

### Custom OCR Languages

Mathpix supports 40+ languages. Specify multiple languages:```bash
--ocr-languages "en,fr,de,es,zh,ja,ko"

Batch Processing

Process multiple PDFs efficiently:

  1. Use pdf://documents/ resource to get complete file list
  2. Filter by date, size, or name patterns
  3. Process each with convert_pdf_to_markdown
  4. Use merge_note with topic_merge strategy for intelligent consolidation

Custom Templates

Create your own note templates by examining the built-in ones:```bash

Get template structure

curl -X POST http://localhost/mcp \
-d '{"method": "tools/call", "params": {"name": "get_note_templates"}}'
```

🤝 Contributing

  1. Fork the repository
  2. Create a feature branch: git checkout -b feature/amazing-feature
  3. Make your changes and add tests
  4. Ensure all tests pass: make test
  5. Commit your changes: git commit -m 'Add amazing feature'
  6. Push to the branch: git push origin feature/amazing-feature
  7. Open a Pull Request

Code Style

  • Follow Go conventions and use gofmt
  • Add tests for new functionality
  • Update documentation for user-facing changes
  • Use structured logging with slog

⚖️ License & Legal Notice

Primary License

This project is licensed under the Apache License 2.0 - see the LICENSE file for details.

🚨 Important: Third-Party Licensing Dependencies

Sibyl uses MuPDF for PDF processing, which has specific licensing requirements:

MuPDF Licensing (AGPL v3)

  • Library: github.com/gen2brain/go-fitz (Go wrapper for MuPDF)
  • Underlying Library: MuPDF (licensed under AGPL v3)
  • Licensor: Artifex Software, Inc.

AGPL v3 Requirements

⚠️ If you use Sibyl in any of the following ways, your entire application MUST be licensed under AGPL v3:

  1. Network Services - Running Sibyl as a web service, API, or SaaS
  2. Modified Distribution - Distributing modified versions of Sibyl
  3. Integration - Incorporating Sibyl into other software
  4. Internal Use - Using Sibyl within an organization over a network

AGPL v3 Obligations

When AGPL applies, you must:
- ✅ License your entire application under AGPL v3 or compatible
- ✅ Provide complete source code to all users
- ✅ Include license notices and copyright information
- ✅ Ensure users can rebuild your application from source

Commercial Licensing Alternative

If you cannot comply with AGPL v3 requirements:

🏢 Contact Artifex Software for a commercial MuPDF license:
- Website: https://mupdf.com/licensing/
- Removes AGPL obligations
- Enables proprietary/commercial deployment
- Required for closed-source applications

Other Dependencies (Permissively Licensed)

All other dependencies use permissive licenses:
- github.com/mark3labs/mcp-go - MIT License
- github.com/joho/godotenv - MIT License
- google.golang.org/api - Apache License 2.0
- Go standard library - BSD-style License

License Compatibility Summary

Use Case License Required Commercial License Needed
Open source project (AGPL v3) ✅ Free ❌ No
Internal company tool ⚠️ AGPL v3 ❌ No*
SaaS/Web service ⚠️ AGPL v3 ❌ No*
Proprietary software ❌ Not possible ✅ Yes
Commercial distribution ❌ Not possible ✅ Yes

*As long as you comply with AGPL v3 source distribution requirements

Recommendation for Users

  1. Open Source Projects: Use Sibyl freely under AGPL v3
  2. Commercial Projects: Evaluate if AGPL v3 compliance is feasible
  3. Proprietary Products: Contact Artifex for commercial licensing
  4. Consulting: Consider legal review for complex licensing scenarios

For questions about licensing compliance, consult with a qualified legal professional.

🙏 Acknowledgments

  • MCP Go SDK - MCP server implementation
  • Mathpix - OCR processing service
  • MuPDF - PDF processing and rendering
  • Google Drive API - Document storage and access

常见问题

6 问
答案由本站依据条目字段整理,不引用上游原文
Sibyl 50762f7b 怎么安装?

资料没有给出安装命令,请到官方仓库或文档按说明接入;这类条目多数是生态应用,或尚未发布可安装包。

Sibyl 50762f7b 怎么在 Claude Desktop、Claude、VS Code 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。

Sibyl 50762f7b 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。

Sibyl 50762f7b 需要 API Key 吗?

资料标注为无需认证,装好就能连;如果实际调用时报鉴权错误,按官方文档补上凭据。

Sibyl 50762f7b 支持哪些客户端?

资料标注支持 Claude Desktop、Claude、VS Code;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。

Sibyl 50762f7b 能提供哪些能力?

资料显示能力标签包括 知识管理、PDF处理、Go、Claude、文档转换。具体工具清单以官方文档为准。

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 2026-05-17。