PDF.co MCP. Extract structured data and convert documents from chat.

Listen up. This isn't some basic converter you use when you're bored. The PDF.co server gives your AI client a full suite of tools to handle documents—it’s for parsing, converting, merging, and locking down PDFs right in the chat window. You'll use it anytime you need your agent to deal with messy files and turn them into clean, actionable data without you having to copy-paste a single thing.

Converting Documents to Structured Data
You can make your AI client read every kind of PDF structure using dedicated conversion tools. If the document has tables, use pdf_to_csv and it'll spit out a perfect Comma Separated Values file you can use in any spreadsheet program. For deeper data analysis, run pdf_to_json to extract and structure all the content into a standardized JSON object that your code can actually read.

If you need something more rigid, pdf_to_xml converts the entire document's structure into an XML file format. Even if all you need is raw reading material, pdf_to_text handles converting the whole PDF down to simple, clean plain text.

Extracting Specific Data and Handling Scans
Sometimes you don't want the whole thing; you just need pieces of info. You can pull basic document information—like who created it, what the title is, or when it was made—by running extract_pdf_meta. If you've got a scanned invoice or some old paperwork that isn't digital text, don't sweat it.

Use ocr_image to run Optical Character Recognition on an uploaded image; it extracts usable text even if the original document was just ink on paper. For massive PDFs, if you only need sections three through five, use split_pdf, and it’ll cut that one big file into several smaller parts for you.

Combining and Managing Files
When your workflow requires multiple inputs, this server handles the heavy lifting. You can run merge_pdfs to take two or more separate PDF documents—say, quarterly reports from different departments—and combine them into a single output file. On the flip side of organization, you might need to mess with security.

If a document is locked down and you need access, use unprotect_pdf to strip away existing passwords so your agent can work on it. Conversely, if you're sending something sensitive, you can run protect_pdf to add password protection, restricting who can view or edit the file.

Utility and Monitoring
Your AI client keeps track of everything running in the background. When a big job—like converting 50 files—is queued up, use check_job_status to see exactly where that document processing is at. Plus, you can keep an eye on your usage with get_account_info, which pulls up your current service credit balance and account metrics so you know what's left.

Basically, it gives you the whole damn toolbox for making PDFs into usable data.