extract_text_from_fileLABSCO SUMMARY
A local document folder becomes something you can put questions to.
Reading and writing live in the same server, which is what makes the round trip work: pull the numbers out of a stack of reports, then have the summary written back as a formatted workbook rather than pasted into chat.
What it is
A Python MCP server for local documents. It extracts text from PDF, DOCX, PPTX and XLSX files, one file or a whole directory with optional recursion, and creates Excel workbooks, Word documents and structured reports.
What you get
- Text extracted from a single PDF, DOCX, PPTX or XLSX file
- A whole directory processed, subdirectories included when you ask for them
- The supported files in a directory listed, with their metadata
- Excel workbooks created with multiple sheets and applied formatting
- Word documents created with formatting
- Structured reports generated with summaries, from data you pass in
Requirements
Python 3.8 or higher and the packages in requirements.txt — pdfplumber, openpyxl and pandas among them. It works on paths on the machine running it, and there is nothing to authenticate.
Setup effort
Build from source — git clone https://github.com/mehmetozcan-zz/KnowledgeBaseMCP.git
USE CASES
01
Getting answers out of a folder of reports
List what is readable in the directory, then extract across all of it in one pass.
02
Reading one PDF properly
Extract the text from that file rather than working from a summary of it.
03
Turning findings into a spreadsheet somebody else can open
Create the workbook with the sheets and formatting already applied.
04
Writing the summary up as a document
Produce a formatted Word file, or a structured report with the summary sections filled in.
05
Working out what is in a directory before processing it
List the supported files and their metadata, so a long extraction is not a surprise.
9 tools
Listed in project docs. Your client may show a different list.
extract_text_from_directorylist_supported_filescreate_docx_documentcreate_structured_reportcreate_xlsx_workbookcreate_xlsx_from_dataframeappend_to_xlsxcreate_xlsx_reportBEFORE YOU INSTALL
What to know first
Tools you get
9 tools named in the project's documentation.
Requires
- python on PATH
- Python 3.8 or later
Where it runs
The documented client runs the server on your machine.
INSTALLATION
Connect your client
You clone this one yourself.
Server setup
TERMINAL
git clone https://github.com/mehmetozcan-zz/KnowledgeBaseMCP.git
cd KnowledgeBaseMCP
pip install -r requirements.txtMCP Clients
Claude Desktop
CLAUDE DESKTOP · JSON
{
"mcpServers": {
"knowledgebase": {
"command": "python",
"args": [
"<YOUR_PATH_TO_KNOWLEDGE_BASE_MCP>/launch_mcp.py"
]
}
}
}
