Markpdf — PDF to Markdown API for AI Agents and RAG Pipelines. Convert PDF, DOCX and XLSX to Markdown.
Markpdf is the agent-first gateway that turns heavy PDF, DOCX, XLSX, PPTX, CSV and HTML into clean Markdown. Fast path for text-native files, OCR only when needed, token-aware response headers.
PDF to Markdown, DOCX to Markdown, XLSX, PPTX, CSV and HTML to Markdown converter
Convert any document to Markdown
Markpdf converts PDF, DOCX, XLSX, CSV and PPTX to clean Markdown built for AI agents and RAG pipelines. Pick a format below for details on how each conversion works.
PDF to Markdown
Text-native PDFs on the fast path, OCR fallback for scanned pages.
DOCX to Markdown
Word headings, tables and lists mapped to clean Markdown.
XLSX to Markdown
Excel spreadsheets, every sheet converted to a Markdown table.
CSV to Markdown
Delimiter and quoting detected automatically, no config.
PPTX to Markdown
Slide text and speaker notes, one Markdown section per slide.
An API, not another PDF reader
Markpdf is not a desktop app, browser extension, or PDF reader. It's an API endpoint your backend or AI agent calls to turn PDF, DOCX, XLSX, CSV and PPTX into Markdown — built for RAG pipelines and LLM document ingestion, not for humans clicking buttons.
Markdown built for
AI agents
Markpdf sits between your documents and your agents. A Kong router validates, meters and routes each request: cheap text-native PDFs go to a Cloudflare Edge worker, heavy files to a PyMuPDF / MarkItDown Full Engine. No OCR unless the document actually needs it.
Direct text/markdown responses, raw uploads without multipart overhead, signed-URL streaming up to 500MB, and token-aware headers so the agent can decide cost versus quality. No external LLMs, no storage, no cache.
Pay for conversion, not chaos
Two payment models, zero fixed subscriptions. You choose how to pay when you create your account.
No monthly fee. Top up whenever you want, you choose how much to spend.
Saved card, what you use gets billed automatically. Built for volume.