PDF Extraction MCP
Extract text, metadata, and tables from PDFs via public URL or base64. Agent-friendly, provider-free.
Documentation
PDF Extraction MCP Server
Model Context Protocol server that extracts text, metadata, and tables from PDF documents — helping AI agents read invoices, contracts, research papers, and reports.
Public Streamable HTTP endpoint
http://64.227.2.61/pdf-mcp/mcp
Tools
| Tool | Args | Returns |
|---|---|---|
| extract_pdf | url | base64_data, filename, include_tables | JSON: text, text_by_page, metadata, words, tables |
| extract_pdf_text | url | base64_data, filename | JSON: text + word stats only |
| extract_pdf_tables | url | base64_data, filename | JSON: detected tables (rows/cols/page) |
Input
- url — public http(s) link to a PDF
- base64_data — base64-encoded PDF bytes (up to 5MB)
REST API (same engine)
POST http://64.227.2.61/pdf/api/v1/pdf/extract · text · tables · payments (Solana USDC). Webhook/OpenAPI at /pdf/openapi.json.
Example
extract_pdf_text(url="https://example.com/report.pdf") → {"filename":"report.pdf","source":"url","page_count":3, "text":"full text...","words":{"total":1200,"unique":300}}
Contact: samayhuf@gmail.com