PDF to Markdown MCP — PDF into clean Markdown, with a scan check first
Agent tools · reviewed listing
Converts a PDF into Markdown and tells you up front whether the file is even convertible. Built for agents and pipelines that have to read documents they did not create: RAG ingestion, document review, translation and DTP work. Connect over Streamable HTTP at https://pdf2md.dudko.dev/mcp — the endpoint is OAuth-protected, so an unauthenticated call answers 401 with a WWW-Authenticate challenge. Two things make it unusual. First, there is no OCR and it says so loudly: a scanned PDF answers 422 naming the type rather than returning plausible nonsense, and a detect call classifies text, scan or mix in milliseconds so a caller can route before paying for anything. Second, a json response reports page count, detected type, layout and which pages would come back thin on text — useful as a routing signal, though it flags sparse pages such as title pages too, so it is a text-density measure rather than a scan verdict. A compact profile exists specifically to spend fewer LLM tokens; a zip response returns the Markdown plus extracted images exactly as linked, with vector figures as SVG carrying the page text. Parsed in-process by a WebAssembly module: uploads are held in memory for the length of the request, never written to disk. Up to 32 MB or 300 pages, whichever bites first. Free allowance on every account, paid plans beyond it. The same parser also runs fully in the browser inside Claude on iOS and claude.ai, where the file is never uploaded (verified 2026-10-05).
Agent tools reviewed 33 opens
Listing owners: outbound opens and referrers are on your private submission status bookmark (the thank-you / status link you saved when you submitted) — open the Analytics tab. That dashboard is only yours; the public page only shows the open count badge.