files-vector-mcp-server 0.1.0

A FastMCP server for file vectorization and similarity search

files-vector-mcp

Note: This project is entirely generated by coze

A FastMCP server for file vectorization and similarity search with automatic directory monitoring.

Features

• 🔍 Auto-monitor directories and process new/changed files
• 📄 Support multiple file types (Markdown, PDF, DOCX, PPTX, images)
• 🧩 Intelligent chunking with overlap for long documents
• 🚀 Fast vector search with PostgreSQL+pgvector
• ⚡ UVX-compatible entry point for quick startup

Installation

With pip

pip install files-vector-mcp-server

With uvx (recommended)

uvx files-vector-mcp-server

Configuration

Environment

# OpenAI Configuration
OPENAI_API_KEY=your_openai_api_key
OPENAI_API_URL=https://api.openai.com/v1
EMBEDDING_MODEL=text-embedding-ada-002
EMBEDDING_DIM=1536

# Database Configuration
DB_CONNECTION_STRING=postgres://user:password@localhost:5432/vector_db

# Monitoring Configuration
WATCH_TOPICS={"docs": ["./docs"], "notes": ["./notes"]}

# Optional MinerU API (for advanced OCR and document processing)(Optional & Untested)
MINERU_API_KEY=your_mineru_key
MINERU_API_URL=https://mineru.net/api/v4

# Chunk Configuration
CHUNK_SIZE=4000
CHUNK_OVERLAP=200

Usage

Basic Usage

files-vector-mcp

With Custom Config

files-vector-mcp --config path/to/your/config.toml

File Type Support

File Type	Handling Method	Requirements
Markdown (.md)	Direct text extraction	-
PDF	MinerU API or PyPDF2	MinerU API or PyPDF2
DOCX	python-docx	python-docx
PPTX	python-pptx	python-pptx
Images	MinerU OCR	MinerU API
Text (.txt)	Direct text extraction	-

Project Structure

src/files-vector-mcp-server/
├── __init__.py
├── __main__.py       # Entry point
├── main.py           # Main application
├── config/           # Configuration
├── services/         # Business logic
└── utils/            # Helper functions

Development

# Clone repository
git clone https://github.com/yourusername/files-vector-mcp.git
cd files-vector-mcp

# Install dependencies
uv venv
source .venv/bin/activate
pip install -e .[dev]

License

Apache 2.0 License - see LICENSE file for details.