coffee_mcp_server

coffee_mcp_server

0
file_systemsai_chatbotdatabases

Coffee MCP Server is a document extraction and processing API built with FastAPI, enabling asynchronous text and table extraction and embedding generation from documents. It utilizes OCR technology for efficient document handling and supports MongoDB for job tracking.

Coffee MCP Server

A powerful document extraction and processing API server built with FastAPI. This server is designed to extract text, tables, and generate embeddings from documents using advanced OCR techniques. It handles documents asynchronously, making it suitable for processing large files without blocking the client.

Key Features

  • Asynchronous document processing
  • Page-by-page PDF processing
  • Text extraction using OCR
  • Table detection and extraction
  • Generation of text embeddings
  • MongoDB storage for job tracking
  • RESTful API with comprehensive endpoints
  • Background thread processing

Setup Guide

Prerequisites

  • Python 3.8+
  • MongoDB
  • Tesseract OCR engine
  • Poppler
  • API keys for embedding providers (OpenAI, Anthropic)

Installation Steps

  1. Clone the repository and navigate to coffee_mcp_server.
  2. Install system and Python dependencies.
  3. Configure environment variables in a .env file.
  4. Start MongoDB and run the server with uvicorn.

API Endpoints

Document Extraction API

  • POST /v1/extract_data: Submit a document for extraction.
  • GET /v1/extract_data_job: Check the status of a document extraction job.
  • GET /v1/extract_data_result: Retrieve the result of a completed document extraction job.

Architecture

Coffee MCP Server comprises components such as the FastAPI application, API routes, document processor, text extractor, format handlers, and embedding generator. It uses MongoDB for storage and supports real-time updates for document processing.