PDF To Text

Local MCP server that extracts text from PDFs, using macOS PDFKit and Vision OCR with an optional AI retry.

What you can do

This server gives an assistant the ability to turn PDF files into text. It reads PDFs from a required root directory and extracts their text locally, falling back to macOS PDFKit and Vision OCR when needed. If local extraction is inadequate, an optional AI retry can be enabled with an OpenAI key. Successful extractions can optionally be saved as text files beside the source PDF.

When to use it

Use it when an assistant needs to read or search the contents of PDF documents on a Mac without uploading them to a cloud service. It is a good fit for local document workflows where OCR of scanned pages matters. Skip it if you need cross-platform support or do not have macOS.

Requirements

Runs as a local stdio process, desktop only. Requires an existing directory set in PDF_TO_TEXT_MCP_ROOT, which defines the filesystem namespace for PDF input and text output. Optional settings include OPENAI_API_KEY for AI retry, OPENAI_MODEL (defaults to gpt-5.4-mini), AI_MAX_OUTPUT_TOKENS, WRITE_TEXT, and extraction queue and parallelism limits.